package ‘astrochron’ · package ‘astrochron’ january 8, 2019 type package title a...
TRANSCRIPT
Package ‘astrochron’January 8, 2019
Type Package
Title A Computational Tool for Astrochronology
Version 0.9
Date 2019-01-08
Author Stephen Meyers [aut,cre], Alberto Malinverno [ctb], Linda Hinnov [ctb], Christian Zee-den [ctb], Vincent Moron [ctb]
Maintainer Stephen Meyers <[email protected]>
Description Routines for astrochronologic testing, astronomical time scale construction, and time se-ries analysis. Also included are a range of statistical analysis and modeling routines that are rele-vant to time scale development and paleoclimate analysis.
Imports multitaper, IDPmisc, fields, doParallel, parallel, iterators,foreach
License GPL-3
NeedsCompilation yes
Repository CRAN
Date/Publication 2019-01-08 19:40:03 UTC
R topics documented:astrochron-package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4anchorTime . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6ar1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6ar1etp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7arcsinT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9armaGen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9asm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10autoPlot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12bandpass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13bergerPeriods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14calcPeriods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15cb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15clipIt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1
2 R topics documented:
confAdjust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17constantSedrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19cosTaper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20delPts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21demean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22detrend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22diffAccum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23divTrend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23dpssTaper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24eAsm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25eAsmTrack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26eha . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27eTimeOpt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29eTimeOptTrack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31etp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32extract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34flip . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35freq2sedrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35gausTaper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36getColor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36getData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37getLaskar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37hannTaper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38headn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39hilbert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39idPts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40imbrie . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41integratePower . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43iso . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45linage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45linterp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47logT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47lowpass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48lowspec . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49makeNoise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52modelA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53mtm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53mtmAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55mtmML96 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56mtmPL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59multiTest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60mwCor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62mwin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64mwStats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65noKernel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66noLow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67pad . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
R topics documented: 3
peak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68periodogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68pl . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71plotEha . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72plS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73prewhiteAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73prewhiteAR1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74pwrLaw . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75pwrLawFit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76rankSeries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77read . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77readMatrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78repl0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79replEps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80resample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80rmNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81sedRamp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82sedrate2time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82slideCor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83sortNave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85stepHeat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85strats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88surrogateCor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88surrogates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90synthStrat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91taner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92testBackground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94testPrecession . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96testTilt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98timeOpt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100timeOptMCMC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103timeOptPlot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106timeOptSim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108timeOptTemplate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110timeOptTemplatePlot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113timeOptTemplateSim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115tones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118traceFreq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119tracePeak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120trackFreq . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121trackPeak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122trim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123trimAT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124trough . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125tune . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125writeCSV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126writeT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4 astrochron-package
wtMean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127xplot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129zoomIn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Index 131
astrochron-package astrochron: A Computational Tool for Astrochronology
Description
This software provides routines for astrochronologic testing, astronomical time scale construction,and time series analysis. Also included are a range of statistical analysis and modeling routines thatare relevant to time scale development and paleoclimate analysis.
Details
Package: astrochronType: PackageVersion: 0.9Date: 2019-01-08License: GPL-3
Note
Please note that this software is undergoing BETA TESTING.
Development of the ’astrochron’ package is partially supported by the U.S. National Science Foun-dation:
CAREER: Deciphering the Beat of a Timeless Rhythm - The Future of Astrochronology (EAR1151438 to S. Meyers)
Collaborative Research: Evolution of the Climate Continuum- Late Paleogene to Present (OCE1003603 to S. Meyers and L. Hinnov)
TO CITE THIS PACKAGE IN PUBLICATIONS, PLEASE USE:
Meyers, S.R. (2014). Astrochron: An R Package for Astrochronology. https://cran.r-project.org/package=astrochron
Also cite the original research papers that document the relevant algorithms, as referenced on thehelp pages for specific functions.
Author(s)
Stephen Meyers
Maintainer: Stephen Meyers <[email protected]>
astrochron-package 5
Examples
### EXAMPLES OF SOME FUNCTIONS AVAILABLE IN THIS SOFTWARE:
### This demo will use a model (series are usually read using the function 'read').data(modelA)
### Interpolate the model stratigraphic series to its median sampling intervalmodelAInterp=linterp(modelA)
### Calculate MTM spectrum using 2pi Slepian tapers, include AR1 condfidence level estimates,### plot power with linear scalemtm(modelAInterp,tbw=2,ar=TRUE,pl=2)
### Perform Evolutive Harmonic Analysis using 2pi Slepian tapers, a window of 8 meters,### pad to 1000 points, and output Harmonic F-test confidence level resultsfCL=eha(modelAInterp,win=8,pad=1000,output=4)
### Extract Harmonic F-test spectrum at approximately 22 meters heightspec=extract(fCL,22)### In this extracted spectrum, identify F-test peak maxima exceeding 90% confidence levelfreqs=peak(spec,level=0.9)[2]### Conduct ASM testing on these peaks# set Rayleigh frequency in cycles/mrayleigh=0.1245274# set Nyquist frequency in cycles/mnyquist=6.66597# set orbital target in 1/kytarget=c(1/405.47,1/126.98,1/96.91,1/37.66,1/22.42,1/18.33)# execute ASMasm(freq=freqs,target=target,rayleigh=rayleigh,nyquist=nyquist,sedmin=0.5,sedmax=3,numsed=100,
linLog=1,iter=100000,output=FALSE)
# Check to see if this is an interactive R session, for compliance with CRAN standards.# YOU SHOULD SKIP THE FOLLOWING LINE IF YOU ARE USING AN INTERACTIVE SESSION.if(interactive()) {
### Interactively track obliquity term in EHA harmonic F-test confidence level resultsfreqs=trackFreq(fCL,fmin=1.2,fmax=2.4,threshold=0.8)
### Convert the spatial frequencies to sedimentation ratessedrate=freq2sedrate(freqs,period=37.66)
### Convert the sedimentation rate curve to a time-space maptime=sedrate2time(sedrate)
### Tune the stratigraphic series using the time-space mapmodelATuned=tune(modelAInterp,time)
### Interpolate the tuned seriesmodelATunedInterp=linterp(modelATuned)
6 ar1
### Perform Evolutive Harmonic Analysis on the tuned serieseha(modelATunedInterp)
}
anchorTime Anchor a floating astrochronology to a radioisotopic age
Description
Anchor a floating astrochronology to a radioisotopic age. The floating astrochronology is centeredon a given (’floating’) time datum and assigned the ’anchored’ age.
Usage
anchorTime(dat,time,age,timeDir=1,flipOut=F,verbose=T,genplot=T)
Arguments
dat Stratigraphic series. First column should be floating time scale, second columnshould be data value.
time ’Floating’ time datum to center record on. Units should be ka.
age Radioisotopic age (or othwerwise) for anchoring at floating ’time’ datum. Unitsshould be ka.
timeDir Direction of ’floating’ time in input record; 1 = elapsed time towards present; 2= elapsed time away from present
flipOut Flip the output (sort so the ages are presented in decreasing order)? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
ar1 Generate AR(1) surrogates
Description
Generate AR(1) surrogates. Implement shuffling algorithm of Meyers (2012) if desired.
Usage
ar1(npts=1024,dt=1,mean=0,sdev=1,rho=0.9,shuffle=F,nsim=1,genplot=T,verbose=T)
ar1etp 7
Arguments
npts number of time series data points
dt sampling interval
mean mean value for AR1 surrogate series
sdev standard deviation for AR1 surrogate series
rho AR(1) coefficient
shuffle Apply secondary shuffle of Gaussian deviates before AR modeling
nsim Number of AR1 surrogate series to generate
genplot generate summary plots (T or F)
verbose verbose output (T or F)
Details
These simulations use the random number generator of Matsumoto and Nishimura [1998]. If shuffle= T, the algorithm from Meyers (2012, pg. 11) is applied: (1) two sets of random sequences ofthe same length are generated, (2) the first random sequence is then sorted, and finally (3) thepermutation vector of the sorted sequence is used to reorder the second random number sequence.This is done to guard against potential shortcomings in random number generation that are specificto spectral estimation.
References
S.R. Meyers, 2012, Seeing red in cyclic stratigraphy: Spectral noise estimation for astrochronology:Paleoceanography, v. 27, PA3328.
ar1etp AR(1) + ETP simulation routine
Description
Simulate a combined AR(1) + ETP signal, plot spectrum and confidence levels
Usage
ar1etp(etpdat=NULL,nsim=100,rho=0.9,wtAR=1,sig=90,tbw=2,padfac=5,ftest=F,fmax=0.1,speed=0.5,pl=2,graphfile=0)
Arguments
etpdat Eccentricity, tilt, precession astronmical series. First column = time, secondcolumn = ETP. If not entered will use default series from Laskar et al. (2004),spanning 0-1000 kyr.
nsim Number of simulations.
rho AR(1) coefficient for noise modeling.
8 ar1etp
wtAR Multiplicative factor for AR1 noise (1= eqivalent to ETP variance). If < 0, etpsignal will be excluded from the simulations (noise only)
sig Demarcate what confidence level (percent) on plots?
tbw MTM time-bandwidth product.
padfac Pad with zeros to (padfac*npts) points, where npts is the number of data points.
ftest Include MTM harmonic f-test results? (T or F)
fmax Maximum frequency for plotting.
speed Set the amount of time to pause before plotting new graph, in seconds.
pl Plot (1) log frequency-log power or (2) linear frequency-linear power?
graphfile Output a pdf or jpg image of each plot? 0 = no, 1 = pdf, 2 = jpeg. If yes, therewill be no output to screen. Individual graphic files will be produced for eachsimluation, for assembling into a movie.
Details
Note: Setting wtAR=1 will provide equal variance contributions from the etp model and the ar1model. More generally, set wtAR to the square root of the desired variance contribution (wtAR=0.5will generate an AR1 model with variance that is 25% of the etp model). If you would like toexclusively evaluate the noise (no etp), set wtAR < 0.
Note: You may use the function etp to generate eccentricity-tilt-precession models.
References
Laskar, J., Robutel, P., Joutel, F., Gastineau, M., Correia, A.C.M., Levrard, B., 2004, A long termnumerical solution for the insolation quantities of the Earth: Astron. Astrophys., Volume 428,261-285.
See Also
getLaskar, and etp
Examples
# run simulations using the default settingsar1etp()
# compare with a second model:# generate etp model spanning 0-2000 ka, with sampling interval of 5 ka.ex1=etp(tmin=0,tmax=2000,dt=5)# run simulations, with rho=-.7, and scaling noise to have 50% of the etp model variancear1etp(etpdat=ex1,rho=0.7,wtAR=sqrt(0.5))
arcsinT 9
arcsinT Arcsine transformation of stratigraphic series
Description
Arcsine transformation of stratigraphic series
Usage
arcsinT(dat,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for arcsine transformation. Input can have any number ofcolumns desired. If two or more columns are input, the first column must belocation (e.g., depth), while remaining columns are data values for transforma-tion.
genplot Generate summary plots? (T or F). This is automatically deactivated if morethan one variable is transformed.
verbose Verbose output? (T or F)
See Also
demean, detrend, divTrend, logT, prewhiteAR, and prewhiteAR1
armaGen Generate autoregressive moving-average model
Description
Generate an autoregressive moving-average time series model
Usage
armaGen(npts=1024,dt=1,m=0,std=1,rhos=c(0.9),thetas=c(0),genplot=T,verbose=T)
Arguments
npts Number of time series data points.
dt Sampling interval.
m Mean value of final time series.
std Standard deviation of final time series.
rhos Vector of AR coefficients for each order.
10 asm
thetas Vector of MA coefficients for each order.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
asm Average Spectral Misfit
Description
Calculate Average Spectral Misfit with Monte Carlo spectra simulations, as updated in Meyers etal. (2012).
Usage
asm(freq,target,fper=NULL,rayleigh,nyquist,sedmin=1,sedmax=5,numsed=50,linLog=1,iter=100000,output=F,genplot=T)
Arguments
freq A vector of candidate astronomical cycles observed in your data spectrum (cy-cles/m). Maximum allowed is 500.
target A vector of astronomical frequencies to evaluate (1/ka). These must be in orderof increasing frequency (e.g., e1,e2,e3,o1,o2,p1,p2). Maximum allowed is 50frequencies.
fper A vector of uncertainties on each target frequency (1/ka). Values should be from0-1, representing uncertainty as a percent of each target frequency. The order ofthe uncertainties must follow that of the target vector. By default, no uncertaintyis assigned.
rayleigh Rayleigh frequency (cycles/m).
nyquist Nyquist frequency (cycles/m).
sedmin Minimum sedimentation rate for investigation (cm/ka).
sedmax Maximum sedimentation rate for investigation (cm/ka).
numsed Number of sedimentation rates to investigate in ASM optimization grid. Maxi-mum allowed is 500.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log)
iter Number of Monte Carlo simulations for significance testing. Maximum allowedis 100,000.
output Return output as a new data frame? (T or F)
genplot Generate summary plots? (T or F)
asm 11
Details
This function will caculate the Average Spectral Misfit between a data spectrum and astronomicaltarget spectrum, following the approach outlined in Meyers and Sageman (2007), and the improve-ments of Meyers et al. (2012).
Value
A data frame containing: Sedimentation rate (cm/ka), ASM (cycles/ka), Null hypothesis signifi-cance level (0-100 percent), Number of astronomical terms fit.
References
S.R. Meyers and B.B. Sageman, 2007, Quantification of Deep-Time Orbital Forcing by AverageSpectral Misfit: American Journal of Science, v. 307, p. 773-792.
S.R. Meyers, B.B. Sageman and M.A. Arthur, 2012, Obliquity forcing of organic matter accumula-tion during Oceanic Anoxic Event 2: Paleoceanography, 27, PA3212, doi:10.1029/2012PA002286.
See Also
eAsm, eAsmTrack, testPrecession, timeOpt, and timeOptSim
Examples
## These frequencies are from modelA (type '?astrochron' for more information).## They are for an 8 meter window, centered at 22 meters height. Units are cycles/m .freq <- c(0.1599833,0.5332776,1.5998329,2.6797201,3.2796575,3.8795948,5.5194235,6.5459830)freq <- data.frame(freq)
## Rayleigh frequency in cycles/mrayleigh <- 0.1245274
## Nyquist frequency in cycles/mnyquist <- 6.66597
## orbital target in 1/ky. Predicted periods for 94 Ma (see Meyers et al., 2012)target <- c(1/405.47,1/126.98,1/96.91,1/37.66,1/22.42,1/18.33)
## percent uncertainty in orbital targetfper=c(0.023,0.046,0.042,0.008,0.035,0.004)
asm(freq=freq,target=target,fper=fper,rayleigh=rayleigh,nyquist=nyquist,sedmin=0.5,sedmax=3,numsed=100,linLog=1,iter=100000,output=FALSE)
12 autoPlot
autoPlot Automatically plot multiple stratigraphic series, with smoothing if de-sired
Description
Automatically plot and smooth specified stratigraphic data, versus location. Data are smoothed witha Gaussian kernel if desired.
Usage
autoPlot(dat,cols=NULL,vertical=T,ydir=NULL,nrows=NULL,plotype=1,smooth=0,xgrid=1,output=F,genplot=T,verbose=T)
Arguments
dat Your data frame; first column should be location identifier (e.g., depth).
cols A vector that identifies the columns to extract (first column automatically ex-tracted).
vertical Generate vertical stratigraphic plots? (T or F) If F, will generate horizontal plots.
ydir Direction for stratigraphic axis in plots (depth,height,time). If vertical=T, then-1 results in values increasing downwards, while 1 results in values increasingupwards. If vertical=F, then -1 results in values increasing toward the left, while1 results in values increasing toward the right.
nrows Number of rows in figure (if vertical = T; otherwise this will be the number ofcolumns).
plotype Type of plot to generate: 1= points and lines, 2 = points, 3 = lines
smooth Width (temporal or spatial dimension) for smoothing with a Gaussian kernel (0= no smoothing); the Gaussian kernel is scaled so that its quartiles (viewed asprobability densities, that is, containing 50 percent of the area) are at +/- 25percent of this value.
xgrid For kernal smoothing: (1) evaluate on ORIGINAL sample grid, or (2) evaluateon EVENLY SPACED grid covering range.
output Output data frame of smoothed values? (T or F)
genplot Generate summary plots (T or F)
verbose Verbose output (T or F)
bandpass 13
bandpass Bandpass filter stratigraphic series
Description
Bandpass filter stratigraphic series using rectangular, Gaussian or tapered cosine (a.k.a. Tukey)window. This function can also be used to notch fiter a record (see examples).
Usage
bandpass(dat,padfac=2,flow=NULL,fhigh=NULL,win=0,alpha=3,p=0.25,demean=T,detrend=F,addmean=T,output=1,xmin=0,xmax=Nyq,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for bandpass filtering. First column should be location (e.g.,depth), second column should be data value.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
flow Lowest frequency to bandpass.
fhigh Highest frequency to bandpass.
win Window type for bandpass filter: 0 = rectangular , 1= Gaussian, 2= Cosine-tapered window (a.k.a. Tukey window).
alpha Gaussian window parameter: alpha is 1/stdev, a measure of the width of theDirichlet kernel. Choose alpha >= 2.5.
p Cosine-tapered (Tukey) window parameter: p is the percent of the data seriestapered (choose 0-1).
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
addmean Add mean value to bandpass result? (T or F)
output Output: (1) filtered series, (2) bandpass filter window.
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Value
bandpassed stratigraphic series.
See Also
lowpass, noKernel, noLow, prewhiteAR, prewhiteAR1, and taner
14 bergerPeriods
Examples
# generate example series with periods of 405 ka, 100 ka, 40ka, and 20 ka, plus noiseex=cycles(freqs=c(1/405,1/100,1/40,1/20),end=1000,dt=5,noisevar=.1)
# bandpass precession term using cosine-tapered windowbandpass_ex <- bandpass(ex,flow=0.045,fhigh=0.055,win=2,p=.4)
# notch filter (remove) obliquity term using cosine-tapered window# if you'd like the final notch filtered record to be centered on the mean proxy# value, set addmean=FALSEnotch_ex <- bandpass(ex,flow=0.02,fhigh=0.03,win=2,p=.4,addmean=FALSE)notch_ex[2] <- ex[2]-notch_ex[2]pl(2)plot(ex,type="l",main="Eccentricity+Obliquity+Precession")plot(notch_ex,type="l",main="Following application of obliquity notch filter")
bergerPeriods Obliquity and precession periods of Berger et al. (1992)
Description
Determine the predicted precession and obliquity periods based on Berger et al. (1992). Values aredetermined by piecewise linear interpolation.
Usage
bergerPeriods(age,genplot=T)
Arguments
age Age (millions of years before present)
genplot Generate summary plots? (T or F)
References
A. Berger, M.F. Loutre, and J. Laskar, 1992, Stability of the Astronomical Frequencies Over theEarth’s History for Paleoclimate Studies: Science, v. 255, p. 560-566.
calcPeriods 15
calcPeriods Calculate eccentricity and precession periods in ka, given g and k inarcsec/yr
Description
Calculate eccentricity and precession periods in ka, given g and k in arcsec/yr.
Usage
calcPeriods(g,k,output=1)
Arguments
g Data frame or matrix with columns representing the fundamental frequencies:g1, g2, g3, g4, g5. Frequencies must be in arcsec/yr.
k Data frame or vector with precession constant (frequency). Frequencies must bein arcsec/yr.
output (1) return results as data frame, (2) return results as a numeric vector.
cb Combine multiple vectors
Description
Bind two vectors together and return result as a data frame. Alternatively, extract specified columnsfrom a data frame, bind them together, and return result as a data frame.
Usage
cb(a,b)
Arguments
a first input vector OR a data frame with >1 column.
b second input vector OR if a is a data frame with > 1 column, a list of columnsto bind.
16 clipIt
Examples
# example datasetx<-rnorm(100)dim(x)<-c(10,10)x<-data.frame(x)
# bind two columnscb(x[1],x[5])
# bind five columnscb(x,c(1,2,4,7,9))
clipIt Create non-linear response by clipping stratigraphic series
Description
Create non-linear response by clipping stratigraphic series below a threshold value. Alternatively,mute response below a threshold value using a contant divisor. Both approaches will enhance powerin modulator (e.g., eccentricity) and diminish power the carrier (e.g., precession).
Usage
clipIt(dat,thresh=NULL,clipval=NULL,clipdiv=NULL,genplot=T,verbose=T)
Arguments
dat Stratigraphic series. First column should be location (e.g., depth), second col-umn should be data value.
thresh Clip below what theshold value? By default will clip at mean value.
clipval What number should be assigned to the clipped values? By default, the value ofthresh is used.
clipdiv Clip using what divisor? A typical value is 2. By default, clipdiv is unity.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
confAdjust 17
confAdjust Adjust spectrum confidence levels for multiple comparisons
Description
Adjust spectrum confidence levels for multiple comparisons, using the Bonferroni correction
Usage
confAdjust(spec,npts,dt,tbw=3,ntap=5,flow=NULL,fhigh=NULL,output=T,xmin=df,xmax=NULL,pl=1,genplot=T,verbose=T)
Arguments
spec A data frame with three columns: frequency, power, background power. If8 columns are input, the results are assumed to come from mtm, mtmML96,lowspec or mtmPL. If 9 columns are input, the results are assumed to comefrom periodogram.
npts Number of points in stratigraphic series.
dt Sampling interval of stratigraphic series.
tbw MTM time-bandwidth product.
ntap Number of DPSS tapers to use.
flow Vector of lower bounds for each frequency band of interest. Order must matchfhigh.
fhigh Vector of upper bounds for each frequency band of interest. Order must matchflow.
output Output data frame? (T or F)
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
pl Plotting option (1-4): 1=linear frequency & log power, 2=log frequency &power, 3=linear frequency & power, 4=log frequency & linear power.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Multiple testing is a common problem in the evaluation of power spectrum peaks (Vaughan et al.,2011; Crampton et al., PNAS). To address the issue of multiple testing, a range of approaches havebeen advocated. This function will conduct an assessment using the Bonferroni correction, whichis the simplest, and also the most conservative, of the common approaches (it is overly pessimistic).
If one is exclusively concerned with particular frequency bands a priori (e.g., those associated withMilankovitch cycles), the statistical power of the method can be improved by restricting the analysisto those frequency bands (use options ’flow’ and ’fhigh’).
18 confAdjust
Application of multiple testing corrections does not guarantee that the spectral background is appro-priate. To address this issue, carefully examine the fit of the spectral background, and also conductsimulations with the function testBackground.
References
J.S. Campton, S.R. Meyers, R.A. Cooper, P.M Sadler, M. Foote, D. Harte, 2018, Pacing of Paleozoicmacroevolutionary rates by Milankovitch grand cycles: Proceedings of the National Academy ofSciences, doi:10.1073/pnas.1714342115.
S. Vaughan, R.J. Bailey, and D.G. Smith, 2011, Detecting cycles in stratigraphic data: Spectralanalysis in the presence of red noise. Paleoceanography 26, PA4211, doi:10.1029/2011PA002195.
See Also
testBackground,multiTest,spec.mtm, lowspec, and periodogram
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=.5)ex[2] = ex[2] + noise[2]
# first, let's use mtm with conventional AR1 backgroundspec=mtm(ex,padfac=1,ar1=TRUE,output=1)
# when blindly prospecting for cycles, it is necessary to consider all of the# observed frequencies in the testconfAdjust(spec,npts=200,dt=5,tbw=3,ntap=5,output=FALSE)
# if, a priori, you are only concerned with the Milankovitch frequency bands,# restrict your analysis to those bands (as constrained by available sedimentation# rate estimates and the frequency resolution of the spectrum). in the example below,# the mtm bandwidth resolution is employed to search frequencies nearby the# Milankovitch-target periods.flow=c((1/400)-0.003,(1/100)-0.003,(1/41)-0.003,(1/20)-0.003)fhigh=c((1/400)+0.003,(1/100)+0.003,(1/41)+0.003,(1/20)+0.003)confAdjust(spec,npts=200,dt=5,tbw=3,ntap=5,flow=flow,fhigh=fhigh,output=FALSE)
# now try with the lowspec method. this uses prewhitening, so it has one less data point.spec=lowspec(ex,padfac=1,output=1)flow=c((1/400)-0.003015075,(1/100)-0.003015075,(1/41)-0.003015075,(1/20)-0.003015075)fhigh=c((1/400)+0.003015075,(1/100)+0.003015075,(1/41)+0.003015075,(1/20)+0.003015075)confAdjust(spec,npts=199,dt=5,tbw=3,ntap=5,flow=flow,fhigh=fhigh,output=FALSE)
# for comparison...confAdjust(spec,npts=199,dt=5,tbw=3,ntap=5,output=FALSE)
constantSedrate 19
constantSedrate Apply a constant sedimentation rate model to transform a spatial se-ries to temporal series
Description
Apply a constant sedimentation rate model to transform a spatial series to temporal series.
Usage
constantSedrate(dat,sedrate,begin=0,timeDir=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series. First column should be location (e.g., depth), second col-umn should be data value.
sedrate Sedimentation rate, in same spatial units as dat.begin Time value to assign to first datum.timeDir Direction of floating time in tuned record: 1 = elapsed time increases with
depth/height; -1 = elapsed time decreases with depth/height)genplot Generate summary plots? (T or F)verbose Verbose output? (T or F)
cosTaper Apply cosine taper to stratigraphic series
Description
Apply a "percent-tapered" cosine taper (a.k.a. Tukey window) to a stratigraphic series.
Usage
cosTaper(dat,p=.25,rms=T,demean=T,detrend=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for tapering. First column should be location (e.g., depth),second column should be data value. If no data is identified, will output a 256point taper to evaluate the spectral properties of the window.
p Cosine-tapered window parameter: p is the percent of the data series tapered(choose 0-1). When p=1, this is equivalent to a Hann taper.
rms Normalize taper to RMS=1 to preserve power for white process? (T or F)demean Remove mean from data series? (T or F)detrend Remove linear trend from data series? (T or F)genplot Generate summary plots? (T or F)verbose Verbose output? (T or F)
20 cycles
See Also
dpssTaper, gausTaper, and hannTaper
cycles Generate harmonic model
Description
Make a time series with specified harmonic components and noise
Usage
cycles(freqs=NULL,phase=NULL,amp=NULL,start=0,end=499,dt=1,noisevar=0,genplot=T,verbose=T)
Arguments
freqs Vector with frequencies to model (’linear’ frequencies).
phase Vector with phases for each frequency (phase in radians). Phases are subtracted.
amp Vector with amplitudes for each frequency.
start First time/depth/height for output.
end Last time/depth/height for output.
dt Sampling interval.
noisevar Variance of additive Gaussian noise.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Value
modeled time series.
Examples
## test signal on pg 38 of Choudhury, Shah, and Thornhill (2008)freqs=c(0.12,0.18,0.30,0.42)phase=c(-pi/3,-pi/12,-pi/4,-3*pi/8)amp=c(1,1,1,1)
cycles(freqs,phase,amp,start=0,end=4095,dt=1,noisevar=0.2)
delPts 21
delPts Interactively delete points in plot
Description
Interactively delete points in x,y plot.
Usage
delPts(dat,del=NULL,cols=c(1,2),ptsize=1,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,plotype=1,genplot=T,verbose=T)
Arguments
dat Data frame containing stratigraphic variable(s) of interest. Any number of columnspermitted.
del A vector of indices indicating points to delete. If specified, the interactive plotis disabled.
cols If you are using the graphical interface, which columns would you like to plot?(default = 1 & 2).
ptsize Size of plotted points.
xmin Minimum x-value (column 1) to plot
xmax Maximum x-value (column 1) to plot
ymin Minimum y-value (column 2) to plot
ymax Maximum y-value (column 2) to plot
plotype Type of plot to generate: 1= points and lines, 2 = points, 3 = lines
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
idPts, iso, trim and trimAT
22 detrend
demean Remove mean value from stratigraphic series
Description
Remove mean value from stratigraphic series
Usage
demean(dat,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for mean removal. First column should be location (e.g.,depth), second column should be data value.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
arcsinT, detrend, divTrend, logT, prewhiteAR, and prewhiteAR1
detrend Subtract linear trend from stratigraphic series
Description
Remove linear trend from stratigraphic series
Usage
detrend(dat,output=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for linear detrending. First column should be location (e.g.,depth), second column should be data value.
output 1= output detrended signal; 2= output linear trend
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
arcsinT, demean, divTrend, logT, prewhiteAR, and prewhiteAR1
diffAccum 23
diffAccum Model differential accumulation
Description
Model differential accumulation. The input variable (e.g., insolation, proxy value) is rescaled tosedimentation rate curve varying from sedmin to sedmax. Input series must be evenly sampled intime.
Usage
diffAccum(dat,sedmin=0.01,sedmax=0.02,dir=1,genplot=T,verbose=T)
Arguments
dat Model input series with two columns. First column must be time in ka, secondcolumn should be data value. Data series must be evenly sampled in time.
sedmin Minimum sedimentation rate (m/ka)
sedmax Maximum sedimentation rate (m/ka)
dir 1=peaks have higher accumulation rate, -1=troughs have higher accumulationrate
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Examples
# generate model with one 20 ka cycleex <- cycles(1/20)
diffAccum(ex)
divTrend Divide by linear trend in stratigraphic series
Description
Divide data series value by linear trend observed in stratigraphic series
Usage
divTrend(dat,genplot=T,verbose=T)
24 dpssTaper
Arguments
dat Stratigraphic series for div-trending. First column should be location (e.g.,depth), second column should be data value.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
arcsinT, demean, detrend, logT, prewhiteAR, and prewhiteAR1
dpssTaper Apply DPSS taper to stratigraphic series
Description
Apply a single Discrete Prolate Spheroidal Sequence (DPSS) taper to a stratigraphic series
Usage
dpssTaper(dat,tbw=1,num=1,rms=T,demean=T,detrend=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for tapering. First column should be location (e.g., depth),second column should be data value. If no data is identified, will output a 256point taper to evaluate the spectral properties of the window.
tbw Time-bandwidth product for the DPSS
num Which one of the DPSS would you like to use?
rms Normalize taper to RMS=1 to preserve power for white process? (T or F)
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
cosTaper, gausTaper, and hannTaper
eAsm 25
eAsm Evolutive Average Spectral Misfit
Description
Calculate Evolutive Average Spectral Misfit with Monte Carlo spectra simulations, as updated inMeyers et al. (2012).
Usage
eAsm(spec,siglevel=0.9,target,fper=NULL,rayleigh,nyquist,sedmin=1,sedmax=5,numsed=50,linLog=1,iter=100000,ydir=1,output=4,genplot=F)
Arguments
spec Time-frequency spectral results to evaluate. Must have the following format:column 1=frequency; remaining columns (2 to n)=probability; titles for columns2 to n must be the location (depth or height). Note that this format is ouput byfunction eha.
siglevel Threshold level for filtering peaks.
target A vector of astronomical frequencies to evaluate (1/ka). These must be in orderof increasing frequency (e.g., e1,e2,e3,o1,o2,p1,p2). Maximum allowed is 50frequencies.
fper A vector of uncertainties on each target frequency (1/ka). Values should be from0-1, representing uncertainty as a percent of each target frequency. The order ofthe uncertainties must follow that of the target vector. By default, no uncertaintyis assigned.
rayleigh Rayleigh frequency (cycles/m).
nyquist Nyquist frequency (cycles/m).
sedmin Minimum sedimentation rate for investigation (cm/ka).
sedmax Maximum sedimentation rate for investigation (cm/ka).
numsed Number of sedimentation rates to investigate in ASM optimization grid. Maxi-mum allowed is 500.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log)
iter Number of Monte Carlo simulations for significance testing. Maximum allowedis 100,000.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
output Return output as a new data frame? (0 = nothing, 1 = Ho-SL, 2 = ASM, 3 = #astronomical terms, 4 = everything)
genplot Generate summary plots? (T or F)
26 eAsmTrack
Details
Please see function asm for details.
References
S.R. Meyers and B.B. Sageman, 2007, Quantification of Deep-Time Orbital Forcing by AverageSpectral Misfit: American Journal of Science, v. 307, p. 773-792.
S.R. Meyers, 2012, Seeing Red in Cyclic Stratigraphy: Spectral Noise Estimation for Astrochronol-ogy: Paleoceanography, 27, PA3228, doi:10.1029/2012PA002307.
S.R. Meyers, B.B. Sageman and M.A. Arthur, 2012, Obliquity forcing of organic matter accumula-tion during Oceanic Anoxic Event 2: Paleoceanography, 27, PA3212, doi:10.1029/2012PA002286.
See Also
asm, eAsmTrack, eha, testPrecession, timeOpt, and timeOptSim
Examples
# use modelA as an exampledata(modelA)
# interpolate to even sampling intervalmodelAInterp=linterp(modelA)
# perform EHA analysis, save harmonic F-test confidence level results to 'spec'spec=eha(modelAInterp,win=8,step=2,pad=1000,output=4)
# perform Evolutive Average Spectral Misfit analysis, save results to 'res'res=eAsm(spec,target=c(1/405.47,1/126.98,1/96.91,1/37.66,1/22.42,1/18.33),rayleigh=0.1245274,
nyquist=6.66597,sedmin=0.5,sedmax=3,numsed=100,siglevel=0.8,iter=10000,output=4)
# identify minimum Ho-SL in each record and plotpl(1)eAsmTrack(res[1],threshold=0.05)
# extract Ho-SL result at 18.23 mHoSL18.23=extract(res[1],get=18.23,pl=1)
# extract ASM result at 18.23 masm18.23=extract(res[2],get=18.23,pl=0)
eAsmTrack Track ASM null hypothesis significance level minima in eASM results
Description
Track ASM null hypothesis significance level minima in eASM results.
eha 27
Usage
eAsmTrack(res,threshold=.5,ydir=-1,genplot=T,verbose=T)
Arguments
res eAsm results. Must have the following format: column 1=sedimentation rate;remaining columns (2 to n)=Ho-SL; titles for columns 2 to n must be the location(depth or height). Note that this format is ouput by function eAsm.
threshold Threshold Ho-SL value for analysis and plotting.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Please see function eAsm for details.
See Also
asm, eAsm, and eha
eha Evolutive Harmonic Analysis & Evolutive Power Spectral Analysis
Description
Evolutive Harmonic Analysis & Evolutive Power Spectral Analysis using the Thomson multitapermethod (Thomson, 1982)
Usage
eha(dat,tbw=2,pad,fmin,fmax,step,win,demean=T,detrend=T,siglevel=0.90,sigID=F,ydir=1,output=0,pl=1,palette=1,centerZero=T,ncolors=100,xlab,ylab,genplot=2,verbose=T)
Arguments
dat Stratigraphic series to analyze. First column should be location (e.g., depth),second column should be data value.
tbw MTM time-bandwidth product (<=10)
pad Pad with zeros to how many points? Must not factor into a prime number >23.Maximum number of points is 200,000.
fmin Smallest frequency for analysis and plotting.
fmax Largest frequency for analysis and plotting.
28 eha
step Step size for EHA window, in units of space or time.
win Window size for EHA, in units of space or time.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
siglevel Significance level for peak identification/filtering (0-1)
sigID Identify signficant frequencies on power, amplitude, and probabilty plots. Onlyapplies when one spectrum is calculated. (T or F)
ydir Direction for y-axis in EHA plots (depth,height,time). -1 = values increasedownwards (slower plotting), 1 = values increase upwards
output Return output as new data frame? 0=no; 1=all results; 2=power; 3=amplitude;4=probability; 5=significant frequencies (only for one spectrum); 6=significantfrequencies and their probabilities (only for one spectrum)
pl Plot logarithm of spectral power (1) or linear spectral power (2)?
palette What color palette would you like to use? (1) rainbow, (2) grayscale, (3) blue,(4) red, (5) blue-white-red (if values are negative and positive, white is centeredon zero)
centerZero Center color scale on zero (use an equal number of postive and negative colordivisions)? (T or F)
ncolors Number of colors steps to use in palette.
xlab Label for x-axis. Default = "Frequency"
ylab Label for y-axis. Default = "Location"
genplot Plotting options. 0= no plots; 1= power, amplitude, f-test, probability; 2=dataseries, power, amplitude, probability; 3= data series, power, normalized ampli-tude (maximum in each window normalized to unity), normalized amplitude fil-tered at specified siglevel; 4= data series, normalized power (maximum in eachwindow normalized to unity), normalized amplitude (maximum in each windownormalized to unity), normalized amplitude filtered at specified siglevel
verbose Verbose output? (T or F)
References
Thomson, D. J., 1982, Spectrum estimation and harmonic analysis, Proc. IEEE, 70, 1055-1096,doi:10.1109/PROC.1982.12433.
See Also
extract, lowspec, mtmAR, mtmML96, periodogram, trackFreq and traceFreq
Examples
## as an example, evaluate the modelAdata(modelA)
## interpolate to even sampling interval of 0.075 mex1=linterp(modelA, dt=0.075)
eTimeOpt 29
## perform EHA with a time-bandwidth parameter of 2, using an 7.95 meter window, 0.15 m step,## and pad to 1000 points## set labels for plots (optional)eha(ex1,tbw=2,win=7.95,step=0.15,pad=1000,xlab="Frequency (cycles/m)",ylab="Height (m)")
## for comparison generate spectrum for entire record, using time-bandwidth parameter of 3, and## pad to 5000 points## start by making a new plotpl(1)eha(ex1,tbw=3,win=38,pad=5000,xlab="Frequency (cycles/m)")
eTimeOpt eTimeOpt: Evolutive implementation of TimeOpt (Meyers, 2015; Mey-ers, 2019)
Description
eTimeOpt: Evolutive implementation of TimeOpt (Meyers, 2015; Meyers, 2019).
Usage
eTimeOpt(dat,win=dt*100,step=dt*10,sedmin=0.5,sedmax=5,numsed=100,linLog=1,limit=T,fit=1,fitModPwr=T,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,detrend=T,ydir=1,output=1,genplot=T,check=T,verbose=1)
Arguments
dat Stratigraphic series for astrochronologic assessment. First column should bedepth or height (in meters), second column should be data value.
win Window size, in meters.
step Step size for moving window, in meters.
sedmin Minimum sedimentation rate for investigation (cm/ka).
sedmax Maximum sedimentation rate for investigation (cm/ka).
numsed Number of sedimentation rates to investigate in optimization grid.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log; default value is 1)
limit Limit evaluated sedimentation rates to region in which full target signal can berecovered? (T or F).
fit Test for (1) precession amplitude modulation or (2) short eccentricity amplitudemodulation?
fitModPwr Include the modulation periods in the spectral fit? (T or F)
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka)
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka)
30 eTimeOpt
roll Taner filter roll-off rate, in dB/octave.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with a first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
detrend Remove linear trend from data series? (T or F)
ydir Direction for y-axis in plots (depth,height,time). -1 = values increase down-wards (slower plotting), 1 = values increase upwards
output Which results would you like to return to the console? (0) no output; (1) every-thing, (2) r^2_envelope, (3) r^2_power, (4) r^2_opt
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (0=nothing, 1=minimal, 2=a little more, 3=everything!)
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulations and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, v.30, 1625-1640.
S.R. Meyers, 2019, Cyclostratigraphy and the problem of astrochronologic testing: Earth-ScienceReviews v.190, 190-223.
See Also
tracePeak,trackPeak,timeOpt,timeOptSim, and eTimeOptTrack
Examples
# generate a test signal with precession and eccentricityex=cycles(freqs=c(1/405.6795,1/130.719,1/123.839,1/98.86307,1/94.87666,1/23.62069,1/22.31868,1/19.06768,1/18.91979),end=4000,dt=5)
# convert to meters with a linearly increasing sedimentation rate from 0.01 m/kyr to 0.03 m/kyrex=sedRamp(ex,srstart=0.01,srend=0.03)
# interpolate to median sampling intervalex=linterp(ex)
# evaluate precession & eccentricity power, and precession modulationsres=eTimeOpt(ex,win=20,step=1,fit=1,output=1)
# extract the optimal fits for the power optimizationsedrates=eTimeOptTrack(res[2])
# extract the optimal fits for the envelope*power optimizationsedrates=eTimeOptTrack(res[3])
eTimeOptTrack 31
# you can also interactively track the results using functions 'trackPeak' and 'tracePeak'# evaluate the results from the power optimizationsedrates=tracePeak(res[2])sedrates=trackPeak(res[2])
# evaluate the results from the envelope*power optimizationsedrates=tracePeak(res[3])sedrates=trackPeak(res[3])
# evaluate precession & eccentricity power, and short-eccentricity modulationseTimeOpt(ex,win=20,step=1,fit=2,output=0)
eTimeOptTrack Track eTimeOpt r2 maxima
Description
Track eTimeOpt r2 maxima.
Usage
eTimeOptTrack(res,threshold=0,ydir=-1,genplot=T,verbose=T)
Arguments
res eTimeOpt r2 results. Must have the following format: column 1=sedimenta-tion rate; remaining columns (2 to n)=r2; titles for columns 2 to n must be thelocation (depth or height). Note that this format is ouput by function eTimeOpt.
threshold Threshold r2-value for analysis and plotting.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Please see function eTimeOpt for details.
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulations and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, v.30, 1625-1640.
S.R. Meyers, 2019, Cyclostratigraphy and the problem of astrochronologic testing: Earth-ScienceReviews v.190, 190-223.
32 etp
See Also
timeOpt, and eTimeOpt
Examples
# generate a test signal with precession and eccentricityex=cycles(freqs=c(1/405.6795,1/130.719,1/123.839,1/98.86307,1/94.87666,1/23.62069,1/22.31868,1/19.06768,1/18.91979),end=4000,dt=5)
# convert to meters with a linearly increasing sedimentation rate from 0.01 m/kyr to 0.03 m/kyrex=sedRamp(ex,srstart=0.01,srend=0.03)
# interpolate to median sampling intervalex=linterp(ex)
# evaluate precession & eccentricity power, and precession modulationsres=eTimeOpt(ex,win=20,step=1,fit=1,output=1)
# extract the optimal fits for the power optimizationsedrates=eTimeOptTrack(res[2])
# extract the optimal fits for the envelope*power optimizationsedrates=eTimeOptTrack(res[3])
# you can also interactively track the results using functions 'trackPeak' and 'tracePeak'# evaluate the results from the power optimizationsedrates=trackPeak(res[2])sedrates=tracePeak(res[2])
# evaluate the results from the envelope*power optimization optimizationsedrates=trackPeak(res[3])sedrates=tracePeak(res[3])
etp Generate eccentricity-tilt-precession models
Description
Calculate eccentricity-tilt-precession time series using the theoretical astronomical solutions. Bydefault, the Laskar et al. (2004) solutions will be downloaded. Alternatively, one can specify theastronomical solution.
Usage
etp(tmin=NULL,tmax=NULL,dt=1,eWt=1,oWt=1,pWt=1,esinw=T,solution=NULL,standardize=T,genplot=T,verbose=T)
etp 33
Arguments
tmin Start time (ka before present, J2000) for ETP. Default value is 0 ka, unless thedata frame ’solution’ is specified, in which case the first time datum is used.
tmax End time (ka before present, J2000) for ETP. Default value is 1000 ka, unlessthe data frame ’solution’ is specified, in which case the last time datum is used.
dt Sample interval for ETP (ka). Minimum = 1 ka.
eWt Relative weight applied to eccentricity solution.
oWt Relative weight applied to obliquity solution.
pWt Relative weight applied to precession solution.
esinw Use e*sinw in ETP calculation? (T or F). If set to false, sinw is used.
solution A data frame containing the astronomical solution to use. The data frame musthave four columns: Time (ka, positive and increasing), Precession Angle, Obliq-uity, Eccentricity.
standardize Standardize (subtract mean, divide by standard deviation) precession, obliquityand eccentricity series before applying weight and combining? (T or F)
genplot Generate summary plots? (T or F).
verbose Verbose output? (T or F).
Details
Note: If you plan to repeatedly execute the etp function, it is advisable to download the astronomicalsolution once using the function getLaskar.
Note: It is common practice to construct ETP models that have specified variance ratios (e.g.,1:1:1 or 1:0.5:0.5) for eccentricity, obliquity and precession. In order to construct such models, itis necessary to choose ’standardize=T’, and to set the individual weights (eWt, oWt, pWt) to thesquare root of the desired variance contribution.
Value
Eccentricity + tilt + precession.
References
Laskar, J., Robutel, P., Joutel, F., Gastineau, M., Correia, A.C.M., Levrard, B., 2004, A long termnumerical solution for the insolation quantities of the Earth: Astron. Astrophys., Volume 428,261-285.
Laskar, J., Fienga, A., Gastineau, M., Manche, H., 2011, La2010: A new orbital solution for thelong-term motion of the Earth: Astron. Astrophys., Volume 532, A89.
Laskar, J., Gastineau, M., Delisle, J.-B., Farres, A., Fienga, A.: 2011, Strong chaos induced by closeencounters with Ceres and Vesta: Astron. Astrophys., Volume 532, L4.
See Also
getLaskar
34 extract
Examples
# create an ETP model from 10000 ka to 20000 ka, with a 5 ka sampling interval# this will automatically download the astronomical solutionex=etp(tmin=10000,tmax=20000,dt=5)
# alternatively, download the astronomical solution firstex2=getLaskar()ex=etp(tmin=10000,tmax=20000,dt=5,solution=ex2)
extract Extract record from EHA time-frequency output or eAsm output
Description
Extract record from EHA time-frequency output or eAsm output: Use interactive graphical interfaceto identify record.
Usage
extract(spec,get=NULL,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,h=6,w=4,ydir=1,pl=0,ncolors=100,genplot=T,verbose=T)
Arguments
spec Time-frequency spectral results to evaluate, or alternatively, eAsm results toevaluate. For time-frequency results, must have the following format: column1=frequency; remaining columns (2 to n)=power, amplitude or probability; titlesfor columns 2 to n must be the location (depth or height). Note that this formatis ouput by function eha. For eAsm results, must have the following format:column 1=sedimentation rate; remaining columns (2 to n)=Ho-SL or ASM; titlesfor columns 2 to n must be the location (depth or height).
get Record to extract (height/depth/time). If no value given, graphical interface isactivated.
xmin Minimum frequency or sedimentation rate for PLOTTING.xmax Maximum frequency or sedimentation rate for PLOTTING.ymin Minimum depth/height for PLOTTING.ymax Maximum depth/height for PLOTTING.h Height of plot in inches.w Width of plot in inches.ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards
(slower plotting!), 1 = values increase upwards.pl An option for the color plots (0=do nothing; 1=plot log of value [useful for plot-
ting power], 2=normalize to maximum value [useful for plotting amplitude]).ncolors Number of colors to use in plot.genplot Generate summary plots? (T or F)verbose Verbose output? (T or F)
flip 35
See Also
eha
flip Flip stratigraphic series
Description
Flip the stratigraphic order of your data series (e.g., convert stratigraphic depth series to heightseries, relative to a defined datum.)
Usage
flip(dat,begin=0,genplot=T,verbose=T)
Arguments
dat Stratigraphic series. First column should be location (e.g., depth), second col-umn should be data value.
begin Depth/height value to assign to (new) first stratigraphic datum.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
freq2sedrate Convert record of local spatial frequency (from EHA) to sedimentationrate curve
Description
Convert record of local spatial frequency (from EHA) to sedimentation rate curve
Usage
freq2sedrate(freqs,period=NULL,ydir=1,genplot=T,verbose=T)
Arguments
freqs Data frame containing depth/height in first column (meters) and spatial frequen-cies in second column (cycles/m)
period Temporal period of spatial frequency (ka)
ydir Direction for y-axis in plots (depth,height). -1 = values increase downwards(slower), 1 = values increase upwards
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
36 getColor
gausTaper Apply Gaussian taper to stratigraphic series
Description
Apply a Gaussian taper to a stratigraphic series
Usage
gausTaper(dat,alpha=3,rms=T,demean=T,detrend=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for tapering. First column should be location (e.g., depth),second column should be data value. If no data is identified, will output a 256point taper to evaluate the spectral properties of the window.
alpha Gaussian window parameter: alpha is 1/stdev, a measure of the width of theDirichlet kernel. Larger values decrease the width of data window, reduce dis-continuities, and increase width of the transform. Choose alpha >= 2.5.
rms Normalize taper to RMS=1 to preserve power for white process? (T or F)
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
References
Harris, 1978, On the use of windows for harmonic analysis with the discrete Fourier transform:Proceedings of the IEEE, v. 66, p. 51-83.
See Also
cosTaper, dpssTaper, and hannTaper
getColor Query R for color information
Description
Query R for color information.
Usage
getColor(color)
getData 37
Arguments
color The name of the color you are interested in, in quotes.
getData Download file from astrochron data server
Description
Download data file from astrochron server.
Usage
getData(dat="1262-a*")
Arguments
dat A character string that specifies the data file to download. At present there areeight options: "1262-a*" , "926B-18O", "graptolite", "Xiamaling-CuAl", "607-18O" , "AEB-18O" , "Newark-rank", "CDL-rank", "DVCP2017-18O"
getLaskar Download Laskar et al. (2004, 2011a, 2011b) astronomical solutions
Description
Download Laskar et al. (2004, 2011a, 2011b) astronomical solutions.
Usage
getLaskar(sol="la04",verbose=T)
Arguments
sol A character string that specifies the astronomical solution to download: "la04","la10a","la10b","la10c","la10d","la11","insolation"verbose Verbose output? (T or F)
Details
la04 : three columns containing precession angle, obliquity, and eccentricity of Laskar et al. (2004)la10a : one column containing the la10a eccentricity solution of Laskar et al. (2011a)la10b : one column containing the la10b eccentricity solution of Laskar et al. (2011a)la10c : one column containing the la10c eccentricity solution of Laskar et al. (2011a)la10d : one column containing the la10d eccentricity solution of Laskar et al. (2011a)la11 : one column containing the la11 eccentricity solution of Laskar et al. (2011b; please also cite2011a)insolation : one column containing insolation at 65 deg North (W/m^2) during summer solstice,from Laskar et al. (2004)
38 hannTaper
References
J. Laskar, P. Robutel, F. Joutel, M. Gastineau, A.C.M. Correia, and B. Levrard, B., 2004, A longterm numerical solution for the insolation quantities of the Earth: Astron. Astrophys., Volume 428,261-285.
Laskar, J., Fienga, A., Gastineau, M., Manche, H., 2011a, La2010: A new orbital solution for thelong-term motion of the Earth: Astron. Astrophys., Volume 532, A89.
Laskar, J., Gastineau, M., Delisle, J.-B., Farres, A., Fienga, A.: 2011b, Strong chaos induced byclose encounters with Ceres and Vesta, Astron: Astrophys., Volume 532, L4.
hannTaper Apply Hann taper to stratigraphic series
Description
Apply a Hann (Hanning) taper to a stratigraphic series
Usage
hannTaper(dat,rms=T,demean=T,detrend=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for tapering. First column should be location (e.g., depth),second column should be data value. If no data is identified, will output a 256point taper to evaluate the spectral properties of the window.
rms Normalize taper to RMS=1 to preserve power for white process? (T or F)
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
cosTaper, dpssTaper, and gausTaper
headn 39
headn List column numbers for each variable
Description
Execute ’head’ function, with column numbers indicated for each variable. (useful for functionssuch as ’autopl’)
Usage
headn(dat)
Arguments
dat Your data frame.
hilbert Hilbert transform of stratigraphic series
Description
Calculate instantaneous amplitude (envelope) via Hilbert Transform of stratigraphic series
Usage
hilbert(dat,padfac=2,demean=T,detrend=F,output=T,addmean=F,genplot=T,check=T,verbose=T)
Arguments
dat Stratigraphic series to Hilbert Transform. First column should be location (e.g.,depth), second column should be data value.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
output Return results as new data frame? (T or F)
addmean Add mean value to instantaneous amplitude? (T or F)
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (T or F)
40 idPts
Examples
# generate example series with 3 precession terms and noiseex <- cycles(noisevar=.0004,dt=5)# bandpass precession terms using cosine-tapered windowres_ex <- bandpass(ex,flow=0.038,fhigh=0.057,win=2,p=.4)# hilbert transformhil_ex <- hilbert(res_ex)
idPts Interactively identify points in plot
Description
Interactively identify points in x,y plot.
Usage
idPts(dat1,dat2=NULL,ptsize=1,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,logx=F,logy=F,plotype=1,annotate=1,output=1,verbose=T)
Arguments
dat1 Data frame with one, two or three columns. If one column, dat2 must also bespecified. If three columns, the data frame is assumed to represent a stratigraphicseries, and the first column should be depth, height or time.
dat2 Data frame with one column.
ptsize Size of plotted points.
xmin Minimum x-value (column 1) to plot
xmax Maximum x-value (column 1) to plot
ymin Minimum y-value (column 2) to plot
ymax Maximum y-value (column 2) to plot
logx Plot x-axis using logarithmic scaling? (T or F)
logy Plot y-axis using logarithmic scaling? (T or F)
plotype Type of plot to generate: 1= points and lines, 2 = points, 3 = lines
annotate Annotate plot with text indicating coordinates?: 0=none, 1=annotate above point,2=annotate below point
output Return identified points as a data frame? (0) no, (1) return x and y, (2) return in-dex, x and y. If dat1 contains three columns, option 2 will return index, location,x and y.
verbose Verbose output? (T or F)
See Also
delPts, iso, trim and trimAT
imbrie 41
imbrie Imbrie and Imbrie (1980) ice sheet model
Description
An implementation of the Imbrie and Imbrie (1980) ice sheet model
Usage
imbrie(insolation=NULL,Tm=17,b=0.6,times=NULL,initial=0,burnin=100,standardize=T,output=T,genplot=1,verbose=T)
Arguments
insolation Insolation, in ka (negative for future, positive for past). Default is insolationover the past 1000 ka from 65 deg. North, 21 June.
Tm Vector of mean time constants in ka. Default is 17 ka. The order of the Tmvalues should match vectors b and times.
b Vector of nonlinearity coefficient (a value ranging from 0 to 1). Default is 0.6.The order of the b values should match vectors Tm and times.
times Vector of start times for each Tm and b listed above. Leave as NULL if you onlyneed to model one Tm and b value.
initial Initial value for ice volume, relative to centered record. Default is 0.
burnin Number of points for model burn-in. This is required to achieve stable modelresults. Default is 100 points.
standardize Standardize model output to maximum value of one and minimum value of zero?(T or F)
output Output model results? (T or F)
genplot Generate summary plots? (1) plot insolation and ice volume series, (2) plotanimated insolation, ice volume and phase portrait.)
verbose Verbose output? (T or F)
Details
This function will implement the ice volume model of Imbrie and Imbrie (1980), following theconventions of Paillard et al. (1996).
When using the ’times’ vector, consider the following example:
times= c(500,1000)
Tm=c(15,5)
b=c(0.6,0.3)
In this case, a Tm of 15 (b of 0.6) will be applied to model from 0-500 ka, and a Tm of 5 (b of 0.3)will be applied to model 500-1000 ka.
42 imbrie
References
Imbrie, J., and Imbrie, J.Z., (1980), Modeling the Climatic Response to Orbital Variations: Science,v. 207, p. 943-953.
Lisiecki, L. E., and M. E. Raymo, 2005, A Pliocene-Pleistocene stack of 57 globally distributedbenthic d18O records, Paleoceanography, 20, PA1003, doi:10.1029/2004PA001071.
Paillard, D., L. Labeyrie and P. Yiou, (1996), Macintosh program performs time-series analysis:Eos Trans. AGU, v. 77, p. 379.
Examples
## Not run:# make a very simple forcing (on/off)forcing=cycles(0,end=300)forcing[50:150,2]=1plot(forcing,type="l")
# use this forcing to drive the imbrie ice model# set b=0, Tm = 1imbrie(forcing,b=0,Tm=1,output=F)
# let's view the evolution of the ice sheetimbrie(forcing,b=0,Tm=1,output=F,genplot=2)
# now increase the response timeimbrie(forcing,b=0,Tm=10,output=F,genplot=2)
# now model slow growth, fast decayimbrie(forcing,b=0.5,Tm=10,output=F,genplot=2)
# now make a 100 ka cyclic forcingforcing=cycles(1/100,end=300)imbrie(forcing,b=0,Tm=1,output=F,genplot=2)imbrie(forcing,b=0,Tm=10,output=F,genplot=2)imbrie(forcing,b=0.5,Tm=10,output=F,genplot=2)# show burn-inimbrie(forcing,b=0.5,Tm=10,output=F,genplot=2,burnin=0)
# now examine Malutin Milankovitch's hypothesis: 65 deg N, summer solsticeimbrie(b=0.5,Tm=10,output=F,genplot=2,burnin=900)
# use the ice model output to make a synthetic stratigraphic sectionres=imbrie(b=0.5,T=10,output=T,genplot=1,burnin=100)synthStrat(res,clip=F)
# generate ice model for last 5300 ka, using 65 deg. N insolation, 21 June# allow b and Tm values to change as in Lisiecki and Raymo (2005):insolation=getLaskar("insolation")insolation=iso(insolation,xmin=0,xmax=5300)# b is 0.3 from 5300 to 3000 ka, then linearly increases to 0.6 between 3000 and 1500 ka.# b is 0.6 from 1500 ka to present.set_b=linterp(cb(c(1500,3000),c(0.6,0.3)),dt=1)
integratePower 43
set_b=rbind(set_b,c(5400,0.3))# Tm is 5 ka from 5300 to 3000 ka, then linearly increases to 15 ka between 3000 and 1500 ka.# Tm is 15 ka from 1500 ka to present.set_Tm=linterp(cb(c(1500,3000),c(15,5)),dt=1)set_Tm=rbind(set_Tm,c(5400,5))# now run modelex=imbrie(insolation=insolation,Tm=set_Tm[,2],b=set_b[,2],times=set_b[,1])# time-frequency analysis of model resulteha(ex,fmax=0.1,win=500,step=10,pad=5000,genplot=4,pl=2)
## End(Not run)
integratePower Determine the total power within a given bandwidth
Description
Determine the total power within a given bandwidth, and also the ratio of this power to the totalpower in the spectrum (or up to a specified frequency). If bandwidth is not specified, generateinteractive plots for bandwidth selection. For use with the function eha, integratePower can processspectrograms (time-frequency) or single spectra.
Usage
integratePower(spec,flow=NULL,fhigh=NULL,fmax=NULL,unity=F,f0=T,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,npts=NULL,pad=NULL,ydir=1,ncolors=100,
h=6,w=9,ln=F,genplot=T,verbose=T)
Arguments
spec Spectral results to evaluate. If the data frame contains time-frequency results, itmust have the following format: column 1=frequency; remaining columns (2 ton)=power; titles for columns 2 to n must be the location (depth or height). Notethat this format is ouput by function eha. If the data frame contains one spec-trum, it must have the following format: column 1=frequency, column 2=power.
flow Low frequency cutoff for integration. If flow or fhigh are not specified, interac-tive plotting is activated.
fhigh High frequency cutoff for integration. If flow or fhigh are not specified, interac-tive plotting is activated.
fmax Integrate total power up to this frequency.
unity Normalize spectra such that total variance (up to fmax) is unity. (T of F)
f0 Is f(0) included in the spectra? (T or F)
xmin Minimum frequency for PLOTTING.
xmax Maximum frequency for PLOTTING.
44 integratePower
ymin Minimum depth/height/time for PLOTTING. Only used if processing time-frequencyresults.
ymax Maximum depth/height/time for PLOTTING. Only used if processing time-frequency results.
npts The number of points in the processed time series window. This is needed forproper spectrum normalization.
pad The total padded length of the processed time series window. This is needed forproper spectrum normalization.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards. Only used if processing time-frequency results.
ncolors Number of colors to use in plot. Only used if processing time-frequency results.
h Height of plot in inches.
w Width of plot in inches.
ln Plot natural log of spectral results? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Depending on the normalization used, you may want to preprocess the power spectra prior to inte-gration.
See Also
eha
Examples
# generate etp signal over past 10 Maex=etp(tmax=10000)
# evolutive powerpwr=eha(ex,win=500,fmax=.1,pad=2000,output=2,pl=2)
# integrate power from main obliquity termintegratePower(pwr,flow=0.02,fhigh=0.029,npts=501,pad=2000)
iso 45
iso Isolate data from a specified stratigraphic interval
Description
Isolate a section of a uni- or multi-variate stratigraphic data set for further analysis
Usage
iso(dat,xmin,xmax,col=2,logx=F,logy=F,genplot=T,verbose=T)
Arguments
dat Data frame containing stratigraphic variable(s) of interest. First column must belocation (e.g., depth).
xmin Minimum depth/height/time for isolation. If xmin is not specified, it will beselected using a graphical interface.
xmax Maximum depth/height/time for isolation. If xmax is not specified, it will beselected using a graphical interface.
col If you are using the graphical interface to select xmin/xmax, which columnwould you like to plot? (default = 2).
logx Plot x-axis using logarithmic scaling? (T or F)
logy Plot y-axis using logarithmic scaling? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
delPts, idPts, trim and trimAT
linage Tune stratigraphic series to an astronomical target using graphicalinterface
Description
Tune stratigraphic series to an astronomical target using graphical interface similar to Analyseries’Linage’ routine (Paillard et al, 1996).
Usage
linage(dat,target,extrapolate=F,xmin=NULL,xmax=NULL,tmin=NULL,tmax=NULL,size=1,plotype=1,output=1,genplot=T)
46 linage
Arguments
dat Stratigraphic series for tuning, with two columns. First column is depth/height.
target Astronomical tuning target series. First column is time.
extrapolate Extrapolate sedimentation rates above and below ’tuned’ interval? (T or F)
xmin Minimum height/depth to plot.
xmax Maximum height/depth to plot.
tmin Minimum time value to plot.
tmax Maximum time value to plot.
size Multiplicative factor to increase or decrease size of symbols and fonts.
plotype Type of plot to generate: 1= points and lines, 2 = points, 3 = lines
output Return which of the following? 1 = tuned stratigraphic series; 2 = age controlpoints; 3 = tuned stratigraphic series and age control points
genplot Generate additional summary plots (tuned record, time-space map, sedimenta-tion rates)? (T or F)
References
Paillard, D., L. Labeyrie and P. Yiou, 1996), Macintosh program performs time-series analysis: EosTrans. AGU, v. 77, p. 379.
Examples
# Check to see if this is an interactive R session, for compliance with CRAN standards.# YOU CAN SKIP THE FOLLOWING LINE IF YOU ARE USING AN INTERACTIVE SESSION.if(interactive()) {
# generate example series with 3 precession terms and noise using function 'cycles'# then convert from time to space using sedimentation rate that increases from 1 to 7 cm/kaex=sedRamp(cycles(start=1,end=400, dt=2,noisevar=.00005),srstart=0.01,srend=0.07)
# create astronomical target seriestarg=cycles(start=1,end=400,dt=2)
## manually tunetuned=linage(ex,targ)
## should you need to flip the direction of the astronomical target series, use function 'cb':tuned=linage(ex,cb(targ[1]*-1,targ[2]))
}
linterp 47
linterp Piecewise linear interpolation of stratigraphic series
Description
Interpolate stratigraphic series onto a evenly sampled grid, using piecewise linear interpolation
Usage
linterp(dat,dt,start,genplot=T,check=T,verbose=T)
Arguments
dat Stratigraphic series for piecewise linear interpolation. First column should belocation (e.g., depth), second column should be data value.
dt New sampling interval.
start Start interpolating at what time/depth/height value? By default, the first value ofthe stratigraphic series will be used.
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (T or F)
logT Log transformation of stratigraphic series
Description
Log transformation of stratigraphic series.
Usage
logT(dat,c=0,opt=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for log transformation. Input can have any number of columnsdesired. If two or more columns are input, the first column must be location (e.g.,depth), while remaining columns are data values for transformation.
c Constant to add prior to log transformation. Default = 0.
opt (1) use natural logarithm, (2) use log10. Default = 1.
genplot Generate summary plots? (T or F). This is automatically deactivated if morethan one variable is transformed.
verbose Verbose output? (T or F)
48 lowpass
See Also
arcsinT, demean, detrend, divTrend, prewhiteAR, and prewhiteAR1
lowpass Lowpass filter stratigraphic series
Description
Lowpass filter stratigraphic series using rectangular, Gaussian or tapered cosine window. Thisfunction can also be used to highpass filter a record (see examples).
Usage
lowpass(dat,padfac=2,fcut=NULL,win=0,demean=T,detrend=F,addmean=T,alpha=3,p=0.25,xmin=0,xmax=Nyq,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for lowpass filtering. First column should be location (e.g.,depth), second column should be data value.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
fcut Cutoff frequency for lowpass filtering.
win Window type for bandpass filter: 0 = rectangular , 1= Gaussian, 2= Cosine-tapered window.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
addmean Add mean value to bandpass result? (T or F)
alpha Gaussian window parameter: alpha is 1/stdev, a measure of the width of theDirichlet kernal. Larger values decrease the width of data window, reduce dis-continuities, and increase width of the transform. Choose alpha >= 2.5.
p Cosine-tapered window parameter: p is the percent of the data series tapered(choose 0-1).
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
bandpass, noKernel, noLow, prewhiteAR, prewhiteAR1, and taner
lowspec 49
Examples
# generate example series with periods of 405 ka, 100 ka, 40ka, and 20 ka, plus noiseex=cycles(freqs=c(1/405,1/100,1/40,1/20),end=1000,dt=5,noisevar=.1)
# lowpass filter eccentricity terms using cosine-tapered windowlowpass_ex=lowpass(ex,fcut=.02,win=2,p=.4)
# highpass filter obliquity and precession terms using cosine-tapered window# if you'd like the final notch filtered record to be centered on the mean proxy# value, set addmean=FALSEhighpass_ex=lowpass(ex,fcut=.02,win=2,p=.4,addmean=FALSE)highpass_ex[2] <- ex[2]-highpass_ex[2]pl(2)plot(ex,type="l",main="Eccentricity+Obliquity+Precession")plot(highpass_ex,type="l",main="Obliquity+Precession highpassed signal")
lowspec Robust Locally-Weighted Regression Spectral Background Estimation
Description
LOWSPEC: Robust Locally-Weighted Regression Spectral Background Estimation (Meyers, 2012)
Usage
lowspec(dat,decimate=NULL,tbw=3,padfac=5,detrend=F,siglevel=0.9,setrho,lowspan,b_tun,output=0,CLpwr=T,xmin,xmax,pl=1,sigID=T,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for LOWSPEC. First column should be location (e.g., depth),second column should be data value.
decimate Decimate statigraphic series to have this sampling interval (via piecewise linearinterpolation). By default, no decimation is performed.
tbw MTM time-bandwidth product (2 or 3 permitted)
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
detrend Remove linear trend from data series? This detrending is performed followingAR1 prewhitening. (T or F)
siglevel Significance level for peak identification. (0-1)
setrho Define AR1 coefficient for pre-whitening (otherwise calculated). If set to 0, nopre-whitening is applied.
lowspan Span for LOWESS smoothing of prewhitened signal, usually fixed to 1. If usingvalue <1, the method is overly conservative with a reduced false positive rate.
50 lowspec
b_tun Robustness weight parameter for LOWSPEC. By default, this will be estimatedinternally.
output What should be returned as a data frame? (0=nothing; 1=pre-whitened spectrum+ harmonic F-test CL + LOWSPEC background + LOWSPEC CL + 90%-99%LOWSPEC power levels; 2=sig peaks)
CLpwr Plot LOWSPEC noise confidence levels on power spectrum? (T or F)
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
pl Power spectrum plotting: (1) linear frequency-log spectral power, (2) linearfrequency-linear spectral power (3) log frequency-log spectral power, (4) logfrequency-linear spectral power
sigID Identify significant frequencies on power and probabilty plots? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
LOWSPEC is a ’robust’ method for spectral background estimation, designed for the identificationof potential astronomical signals that are imbedded in red noise (Meyers, 2012). The completealgoritm implemented here is as follows: (1) initial pre-whitening with AR1 filter (default) or otherfilter as appropriate (e.g., see function prewhiteAR), (2) power spectral estimation via the multitapermethod (Thomson, 1982), (3) robust locally weighted estimation of the spectral background usingthe LOWESS-based (Cleveland, 1979) procedure of Ruckstuhl et al. (2001), (4) assignment ofconfidence levels using a Chi-square distribution.
NOTE: If you choose to pre-whiten before running LOWSPEC (rather than using the default AR1pre-whitening), specify setrho=0.
Candidiate astronomical cycles are subsequently identified via isolation of those frequencies thatachieve the required (e.g., 90 percent) LOWSPEC confidence level and MTM harmonic F-test con-fidence level. Allowance is made for the smoothing inherent in the MTM power spectral estimateas compared to the MTM harmonic spectrum. That is, an F-test peak is reported if it achieves therequired MTM harmonic confidence level, while also achieving the required LOWSPEC confidencelevel within +/- half the power spectrum bandwidth resolution. One additional criterion is includedto further reduce the false positive rate, a requirement that significant F-tests must occur on a localpower spectrum high, which is parameterized as occurring above the local LOWSPEC backgroundestimate. See Meyers (2012) for futher information on the algorithm.
In this implementation, the ’robustness criterion’ (’b’ in EQ. 6 of Ruckstuhl et al., 2001) has beenoptimized for 2 and 3 pi DPSS, using a ’span’ of 1. By default the robustness criterion will beestimated. Both ’b’ and the ’span’ can be expliclty set using parameters ’b_tun’ and ’lowspan’.Note that it is permissible to decrease ’lowspan’ from its default value, but this will result in anoverly conservative false positive rate. However, it may be necessary to reduce ’lowspan’ to providean approporiate background fit for some stratigraphic data. Another option is to decimate the dataseries prior to spectral estimation.
lowspec 51
Value
If option 1 is selected, a data frame containing the following is returned: Frequency, Prewhitenedpower, harmonic F-test CL, LOWSPEC CL, LOWSPEC background, 90%-99% LOWSPEC powerlevels. NOTE: as of version 0.8, the order of the columns in the output data frame has been changed,for consistency with functions mtm, mtmML96, and mtmPL.
If option 2 is selected, the ’significant’ frequencies are returned (as described above).
References
W.S. Cleveland, 1979, Locally weighted regression and smoothing scatterplots: Journal of theAmerican Statistical Association, v. 74, p. 829-836.
S.R. Meyers, 2012, Seeing Red in Cyclic Stratigraphy: Spectral Noise Estimation for Astrochronol-ogy: Paleoceanography, 27, PA3228, doi:10.1029/2012PA002307.
A.F. Ruckstuhl, M.P Jacobson, R.W. Field, and J.A. Dodd, 2001, Baseline subtraction using robustlocal regression estimation: Journal of Quantitative Spectroscopy & Radiative Transfer, v. 68, p.179-193.
D.J. Thomson, 1982, Spectrum estimation and harmonic analysis: IEEE Proceedings, v. 70, p.1055-1096.
See Also
eha, mtm, mtmAR, mtmML96, periodogram, rfbaseline, and spec.mtm
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=.5)ex[2] = ex[2] + noise[2]
# LOWSPEC analysispl(1, title="lowspec")lowspec(ex)
# compare to MTM spectral analysis, with conventional AR1 noise testpl(1,title="mtm")mtm(ex,ar1=TRUE)
# compare to ML96 analysispl(1, title="mtmML96")mtmML96(ex)
# compare to amplitudes from ehapl(1,title="eha")eha(ex,tbw=3,win=1000,pad=1000)
52 makeNoise
makeNoise Generate noise surrogates from a theoretical power spectrum
Description
Generate noise surrogates from a theoretical power spectrum.
Usage
makeNoise(S,dt=1,mean=0,sdev=1,addPt=F,nsim=1,genplot=T,verbose=T)
Arguments
S Vector or 1-D data frame containing the theoretical power, from f(0) to theNyquist frequency.
dt Sampling interval for surrogate series
mean Mean value for surrogate series
sdev Standard deviation for surrogate series
addPt Did you add a Nyquist frequency? (T or F)
nsim Number of surrogate series to generate
genplot generate summary plots (T or F)
verbose verbose output (T or F)
Details
These simulations use the random number generator of Matsumoto and Nishimura [1998]. Thealgorithm of Timmer and Konig (1995) is employed to generate surrogates from any arbitrary the-oretical power spectrum. See examples section below.
References
M. Matsumoto, and T. Nishimura, (1998), Mersenne Twister: A 623-dimensionally equidistributeduniform pseudo-random number generator, ACM Transactions on Modeling and Computer Simu-lation, 8, 3-30.
J. Timmer and K. Konig (1995), On Generating Power Law Noise, Astronomy and Astrophysics:v. 300, p. 707-710.
Examples
# create theoretical AR1 spectrum, using rho of 0.8rho=0.8freq=seq(0,.5,by=0.005)Nyq=max(freq)AR1 = (1-(0.8^2)) / ( 1 - (2*0.8*cos(pi*freq/Nyq)) + (0.8^2) )plot(freq,AR1,type="l")
modelA 53
# make noise surrogates from the theoretical AR1 spectrummakeNoise(AR1)
modelA Example stratigraphic model series
Description
Example stratigraphic model series.
Usage
modelA
Format
Height (meters), weight percent CaCO3
mtm Multitaper method spectral analysis
Description
Multitaper method (MTM) spectral analysis (Thomson, 1982)
Usage
mtm(dat,tbw=3,ntap=NULL,padfac=5,demean=T,detrend=F,siglevel=0.9,ar1=T,output=0,CLpwr=T,xmin,xmax,pl=1,sigID=T,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for MTM spectral analysis. First column should be location(e.g., depth), second column should be data value.
tbw MTM time-bandwidth product.
ntap Number of DPSS tapers to use. By default, this is set to (2*tbw)-1.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
siglevel Significance level for peak identification. (0-1)
ar1 Estimate conventional AR(1) noise spectrum and confidence levels? (T or F)
CLpwr Plot AR(1) noise confidence levels on power spectrum? (T or F)
54 mtm
output What should be returned as a data frame? (0=nothing; 1= power spectrum +harmonic CL + AR1 CL + AR1 fit + 90%-99% AR1 power levels (ar1 must beset to TRUE to output AR model results); 2=significant peak frequencies; 3=sig-nificant peak frequencies + harmonic CL; 4=internal variables from spec.mtm).Option 4 is intended for expert users, and should generally be avoided.
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
pl Power spectrum plotting: (1) linear frequency-log spectral power, (2) linearfrequency-linear spectral power (3) log frequency-log spectral power, (4) logfrequency-linear spectral power
sigID Identify significant frequencies on power and probabilty plots? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
If ar1=T, candidiate astronomical cycles are identified via isolation of those frequencies that achievethe required (e.g., 90 percent) "red noise" confidence level and MTM harmonic F-test confidencelevel. Allowance is made for the smoothing inherent in the MTM power spectral estimate as com-pared to the MTM harmonic spectrum. That is, an F-test peak is reported if it achieves the re-quired MTM harmonic confidence level, while also achieving the required red noise confidencelevel within +/- half the power spectrum bandwidth resolution. One additional criterion is includedto further reduce the false positive rate, a requirement that significant F-tests must occur on a localpower spectrum high, which is parameterized as occurring above the local red noise backgroundestimate. See Meyers (2012) for futher information.
References
S.R. Meyers, 2012, Seeing Red in Cyclic Stratigraphy: Spectral Noise Estimation for Astrochronol-ogy: Paleoceanography, 27, PA3228, doi:10.1029/2012PA002307.
Rahim, K.J. and Burr W.S. and Thomson, D.J., 2014, Appendix A: Multitaper R package in "Appli-cations of Multitaper Spectral Analysis to Nonstationary Data", PhD diss., Queen’s Univieristy, pp149-183. http://hdl.handle.net/1974/12584
Thomson, D. J., 1982, Spectrum estimation and harmonic analysis, Proc. IEEE, 70, 1055-1096,doi:10.1109/PROC.1982.12433.
See Also
eha, lowspec, mtmAR, mtmML96, periodogram, and spec.mtm
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=.5)
mtmAR 55
ex[2] = ex[2] + noise[2]
# MTM spectral analysis, with conventional AR1 noise testpl(1,title="mtm")mtm(ex,ar1=TRUE)
# compare to ML96 analysispl(1, title="mtmML96")mtmML96(ex)
# compare to analysis with LOWSPECpl(1, title="lowspec")lowspec(ex)
# compare to amplitudes from ehapl(1,title="eha")eha(ex,tbw=3,win=1000,pad=1000)
mtmAR Intermediate spectrum test of Thomson et al. (2001)
Description
Perform the ’intermediate spectrum test’ of Thomson et al. (2001).
Paraphrased from Thomson et al. (2001): Form an intermediate spectrum by dividing MTM byAR estimate. Choose an order P for a predictor. A variety of formal methods are available in theliterature, but practically, one keeps increasing P (the order) until the range of the intermediatespectrum Si(f) (equation (C4) of Thomson et al., 2001) stops decreasing rapidly as a function of P.If the intermediate spectrum is not roughly white, as judged by the minima, the value of P shouldbe increased.
Usage
mtmAR(dat,tbw=3,ntap=NULL,order=1,method="mle",CItype=1,padfac=5,demean=T,detrend=F,output=1,xmin=0,xmax=Nyq,pl=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for analysis. First column should be location (e.g., depth),second column should be data value.
tbw MTM time-bandwidth product.
ntap Number of DPSS tapers to use. By default, this is set to (2*tbw)-1.
order Order of the AR spectrum.
method AR method ("yule-walker", "burg", "ols", "mle", "yw")
CItype Illustrate (1) one-sided or (2) two-sided confidence intervals on plots
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
56 mtmML96
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
output Output (1) intermediate spectrum and confidence levels, (2) intermediate spec-trum, (3) confidence levels
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
pl Plot logarithm of spectral power (1) or linear spectral power (2)?
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
References
Thomson, D. J., L. J. Lanzerotti, and C. G. Maclennan, 2001, The interplanetary magnetic field:Statistical properties and discrete modes, J. Geophys.Res., 106, 15,941-15,962, doi:10.1029/2000JA000113.
See Also
eha, lowspec, mtm, mtmML96, periodogram, and spec.mtm
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=.5)ex[2] = ex[2] + noise[2]
# MTM spectral analysis, with conventional AR1 noise testpl(1,title="mtmAR")mtmAR(ex)
mtmML96 Mann and Lees (1996) robust red noise MTM analysis
Description
Mann and Lees (1996) robust red noise MTM analysis. This function implements several improve-ments to the algorithm used in SSA-MTM toolkit, including faster AR1 model optimization, andmore appropriate ’edge-effect’ treatment.
Usage
mtmML96(dat,tbw=3,ntap=NULL,padfac=5,demean=T,detrend=F,medsmooth=0.2,opt=1,linLog=2,siglevel=0.9,output=0,CLpwr=T,xmin=0,xmax=Nyq,sigID=T,pl=1,genplot=T,verbose=T)
mtmML96 57
Arguments
dat Stratigraphic series for MTM spectral analysis. First column should be location(e.g., depth), second column should be data value.
tbw MTM time-bandwidth product.
ntap Number of DPSS tapers to use. By default, this is set to (2*tbw)-1.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
medsmooth ML96 median smoothing parameter (1 = use 100% of spectrum; 0.20 = use20%)
opt Optimization method for robust AR1 model estimation (1=Brent’s method:fast,2=Gauss-Newton:fast, 3=grid search:slow)
linLog Optimize AR1 model fit using (1) linear power or (2) log(power)?
siglevel Significance level for peak identification. (0-1)
output What should be returned as a data frame? (0=nothing; 1= power spectrum +harmonic CL + AR1 CL + AR1 fit + 90%-99% AR1 power levels + mediansmoothed spectrum; 2=significant peak frequencies; 3=significant peak frequen-cies + harmonic CL)
CLpwr Plot ML96 AR(1) noise confidence levels on power spectrum? (T or F)
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
sigID Identify significant frequencies on power and probabilty plots? (T or F)
pl Power spectrum plotting: (1) linear frequency-log spectral power, (2) linearfrequency-linear spectral power (3) log frequency-log spectral power, (4) logfrequency-linear spectral power
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
This function conducts the Mann and Lees (1996; ML96) "robust red noise" analysis, with animproved median smoothing approach. The original Mann and Lees (1996) approach applies atruncation of the median smoothing window to include fewer frequencies near the edges of thespectrum; while truncation is required, its implementation in the original method often results in an"edge effect" that can produce excess false positive rates at low frequencies, commonly within theeccentricity-band (Meyers, 2012).
To help address this issue, an alternative median smoothing approach is applied that implementsTukey’s robust end-point rule and symmetrical medians (see the function runmed for details). Nu-merical experiments indicate that this approach produces an approximately uniform false positiverate across the spectrum. It should be noted that the false positive rates are still inflated with thismethod, but they are substantially reduced compared to the original ML96 approach. For example,simulations using rho=0.9 (using identical parameters to those in Meyers, 2012) yield median false
58 mtmML96
positive rates of 1.7%, 7.3% and 13.4%, for the 99%, 95% and 90% confidence levels (respectively).This compares with 4.7%, 11.4% and 17.8% using the original approach (see Table 2 of Meyers,2012).
Candidiate astronomical cycles are identified via isolation of those frequencies that achieve therequired (e.g., 90 percent) "robust red noise" confidence level and MTM harmonic F-test confi-dence level. Allowance is made for the smoothing inherent in the MTM power spectral estimateas compared to the MTM harmonic spectrum. That is, an F-test peak is reported if it achieves therequired MTM harmonic confidence level, while also achieving the required robust red noise con-fidence level within +/- half the power spectrum bandwidth resolution. One additional criterion isincluded to further reduce the false positive rate, a requirement that significant F-tests must occuron a local power spectrum high, which is parameterized as occurring above the local robust rednoise background estimate. See Meyers (2012) for futher information.
NOTE: If the (fast) Brent or Gauss-Newton methods fail, use the (slow) grid search approach.
This version of the ML96 algorithm was first implemented in Patterson et al. (2014).
References
Mann, M.E., and Lees, J.M., 1996, Robust estimation of background noise and signal detection inclimatic time series, Clim. Change, 33, 409-445.
Meyers, S.R., 2012, Seeing red in cyclic stratigraphy: Spectral noise estimation for astrochronol-ogy, Paleoceanography, 27, PA3228.
M.O. Patterson, R. McKay, T. Naish, C. Escutia, F.J. Jimenez-Espejo, M.E. Raymo, M.E., S.R.Meyers, L. Tauxe, H. Brinkhuis, and IODP Expedition 318 Scientists,2014, Response of the EastAntarctic Ice Sheet to orbital forcing during the Pliocene and Early Pleistocene, Nature Geoscience,v. 7, p. 841-847.
Thomson, D. J., 1982, Spectrum estimation and harmonic analysis, Proc. IEEE, 70, 1055-1096,doi:10.1109/PROC.1982.12433.
http://www.meteo.psu.edu/holocene/public_html/Mann/tools/MTM-RED/
Tukey, J.W., 1977, Exploratory Data Analysis, Addison.
See Also
eha, lowspec, mtm, mtmAR, periodogram, runmed, and spec.mtm
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=0.5)ex[2] = ex[2] + noise[2]
# run ML96 analysispl(1, title="mtmML96")mtmML96(ex)
# compare to analysis with conventional AR1 noise test
mtmPL 59
pl(1,title="mtm")mtm(ex)
# compare to analysis with LOWSPECpl(1, title="lowspec")lowspec(ex)
# compare to amplitudes from ehapl(1,title="eha")eha(ex,tbw=3,win=1000,pad=1000)
mtmPL Multitaper Method Spectral Analysis with Power Law (1/f) fit
Description
Multitaper Method (MTM) Spectral Analysis with Power Law (1/f) fit
Usage
mtmPL(dat,tbw=3,ntap=NULL,padfac=5,demean=T,detrend=F,siglevel=0.9,flow=NULL,fhigh=NULL,output=0,CLpwr=T,xmin=0,xmax=Nyq,pl=1,sigID=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for MTM spectral analysis. First column should be location(e.g., depth), second column should be data value.
tbw MTM time-bandwidth product.
ntap Number of DPSS tapers to use. By default, this is set to (2*tbw)-1.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
siglevel Significance level for peak identification.
flow Lowest frequency to include in 1/f fit
fhigh Highest frequency to include in 1/f fit
output What should be returned as a data frame? (0=nothing; 1=spectrum + CLs +power law fit; 2=sig peak freqs; 3=sig peak freqs + prob; 4=all)
CLpwr Plot power law noise confidence levels on power spectrum (in addition to thepower law fit)? (T or F)
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
pl Power spectrum plotting: (1) linear frequency-log spectral power, (2) linearfrequency-linear spectral power (3) log frequency-log spectral power, (4) logfrequency-linear spectral power
60 multiTest
sigID Identify signficant frequencies on power and probabilty plots? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Candidiate astronomical cycles are identified via isolation of those frequencies that achieve therequired (e.g., 90 percent) power law confidence level and MTM harmonic F-test confidence level.Allowance is made for the smoothing inherent in the MTM power spectral estimate as compared tothe MTM harmonic spectrum. That is, an F-test peak is reported if it achieves the required MTMharmonic confidence level, while also achieving the required power law confidence level within+/- half the power spectrum bandwidth resolution. One additional criterion is included to furtherreduce the false positive rate, a requirement that significant F-tests must occur on a local powerspectrum high, which is parameterized as occurring above the local red noise background estimate.See Meyers (2012) for futher information.
References
S.R. Meyers, 2012, Seeing Red in Cyclic Stratigraphy: Spectral Noise Estimation for Astrochronol-ogy: Paleoceanography, 27, PA3228, doi:10.1029/2012PA002307.
Rahim, K.J. and Burr W.S. and Thomson, D.J., 2014, Appendix A: Multitaper R package in "Appli-cations of Multitaper Spectral Analysis to Nonstationary Data", PhD diss., Queen’s Univieristy, pp149-183. http://hdl.handle.net/1974/12584
Thomson, D. J., 1982, Spectrum estimation and harmonic analysis, Proc. IEEE, 70, 1055-1096,doi:10.1109/PROC.1982.12433.
See Also
eha, lowspec, mtm, mtmAR, mtmML96, periodogram, and spec.mtm
multiTest Adjust spectral p-values for multiple comparisons
Description
Adjust spectral p-values for multiple comparisons, using a range of approaches.
Usage
multiTest(spec,flow=NULL,fhigh=NULL,pl=T,output=T,genplot=T,verbose=T)
multiTest 61
Arguments
spec A data frame with two columns: frequency, uncorrected confidence level. If8 columns are input, the results are assumed to come from mtm, mtmML96,lowspec or mtmPL. If 9 columns are input, the results are assumed to comefrom periodogram.
flow Vector of lower bounds for each frequency band of interest. Order must matchfhigh.
fhigh Vector of upper bounds for each frequency band of interest. Order must matchflow.
pl Include graphs of uncorrected p-values? (T or F)
output Return results as new data frame? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Multiple testing is a common problem in the evaluation of power spectrum peaks (Vaughan et al.,2011; Crampton et al., PNAS). To address the issue of multiple testing, a range of approaches havebeen advocated. This function will conduct an assessment using six approaches: Bonferroni, Holm(1979), Hochberg (1998), Hommel (1988), Benjamini & Hochberg (1995, a.k.a. "false discoveryrate"), Benjamini & Yekutieli (2001, a.k.a. "false discovery rate"). See the function p.adjust foradditional information on these six approaches.
In conducting these assessments, it is important that the spectral analysis is conducted without zero-padding. If one is exclusively concerned with particular frequency bands a priori (e.g., those asso-ciated with Milankovitch cycles), the statistical power of the method can be improved by restrictingthe analysis to those frequency bands (use options ’flow’ and ’fhigh’).
Application of these multiple testing corrections does not guarantee that the spectral backgroundis appropriate. To address this issue, carefully examine the fit of the spectral background, and alsoconduct simulations with the function testBackground.
References
Y. Benjamini, and Y. Hochberg, 1995, Controlling the false discovery rate: a practical and powerfulapproach to multiple testing. Journal of the Royal Statistical Society Series B 57, 289-300.
Y. Benjamini, and D. Yekutieli, 2001, The control of the false discovery rate in multiple testingunder dependency. Annals of Statistics 29, 1165-1188.
J.S. Campton, S.R. Meyers, R.A. Cooper, P.M Sadler, M. Foote, D. Harte, 2018, Pacing of Paleozoicmacroevolutionary rates by Milankovitch grand cycles: Proceedings of the National Academy ofSciences, doi:10.1073/pnas.1714342115.
S. Holm, 1979, A simple sequentially rejective multiple test procedure. Scandinavian Journal ofStatistics 6, 65-70.
G. Hommel, 1988, A stagewise rejective multiple test procedure based on a modified Bonferronitest. Biometrika 75, 383-386.
Y. Hochberg, 1988, A sharper Bonferroni procedure for multiple tests of significance. Biometrika75, 800-803.
62 mwCor
J.P. Shaffer, 1995, Multiple hypothesis testing. Annual Review of Psychology 46, 561-576. (Anexcellent review of the area.)
S. Vaughan, R.J. Bailey, and D.G. Smith, 2011, Detecting cycles in stratigraphic data: Spectralanalysis in the presence of red noise. Paleoceanography 26, PA4211, doi:10.1029/2011PA002195.
See Also
p.adjust,testBackground,confAdjust,lowspec, mtm, mtmML96, mtmPL, and periodogram
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=.5)ex[2] = ex[2] + noise[2]
# first, let's look at mtm with conventional AR1 backgroundspec=mtm(ex,padfac=1,ar1=TRUE,output=1)
# when blindly prospecting for cycles, it is necessary to consider all of the# observed frequencies in the testmultiTest(cb(spec,c(1,4)),output=FALSE)
# if, a priori, you are only concerned with the Milankovitch frequency bands,# restrict your analysis to those bands (as constrained by available sedimentation# rate estimates and the frequency resolution of the spectrum). in the example below,# the mtm bandwidth resolution is employed to search frequencies nearby the# Milankovitch-target periods.flow=c((1/400)-0.003,(1/100)-0.003,(1/41)-0.003,(1/20)-0.003)fhigh=c((1/400)+0.003,(1/100)+0.003,(1/41)+0.003,(1/20)+0.003)multiTest(cb(spec,c(1,4)),flow=flow,fhigh=fhigh,output=FALSE)
# now try with the lowspec method. this uses prewhitening, so it has one less data point.spec=lowspec(ex,padfac=1,output=1)flow=c((1/400)-0.003015075,(1/100)-0.003015075,(1/41)-0.003015075,(1/20)-0.003015075)fhigh=c((1/400)+0.003015075,(1/100)+0.003015075,(1/41)+0.003015075,(1/20)+0.003015075)multiTest(cb(spec,c(1,4)),flow=flow,fhigh=fhigh,output=FALSE)
# for comparison...multiTest(cb(spec,c(1,4)),output=FALSE)
mwCor Calculate moving window correlation coefficient for two stratigraphicseries, using a ’dynamic window’
mwCor 63
Description
Calculate moving window correlation coefficient for two stratigraphic series, using a ’dynamic win-dow’. This routine adjusts the number of data points in the window so it has a constant duration intime or space, for use with unevenly sampled data.
Usage
mwCor(dat,cols=NULL,win=NULL,conv=1,cormethod=1,output=T,pl=1,genplot=T,verbose=T)
Arguments
dat Your data frame containing stratigraphic data; any number of columns (vari-ables) are permitted, but the first column should be a location identifier (e.g.,depth, height, time).
cols A vector that identifies the two variable columns to be extracted (first columnautomatically extracted).
win Moving window size in units of space or time.
conv Convention for window placement: (1) center each window on a stratigraphiclevel in ’dat’ (DEFAULT), (2) start with the smallest location datum in ’dat’, (3)start with the largest location datum in ’dat’. For options 2 and 3, the center ofthe window will not necessarily coincide with a measured stratigraphic level in’dat’, but edges of the data set are better preserved.
cormethod Method used for calculation of correlation coefficient (1=Pearson, 2=Spearman,3=Kendall)
output Output results? (T or F)
pl (1) Plot results at center of window, or (2) create "string of points plot" as inSageman and Hollander (1999)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
References
B.B. Sageman and D.H. Hollander, 1999, Cross correlation of paleoecological and geochemicalproxies: A holistic approach to the study of past global change, in E. Barrera and C.C. Johnson,eds., GSA Special Paper 332, p. 365-384.
Examples
# generate example seriesex <- cycles(freqs=c(1/40,1/20),noisevar=.2)
# add second variableex[3] <- cycles(freqs=c(1/40,1/20),noisevar=0.2)[2]
# jitter sampling timesex[1]=ex[1]+rnorm(500,sd=5)# sort
64 mwin
ex = ex[order(ex[1],na.last=NA,decreasing=FALSE),]
# run mwCormwCor(ex,win=50)
mwin Determine ’dynamic moving window’ for stratigraphic series, adjust-ing for changing sample density to maintain a window of constantduration
Description
Determine start and end points for a moving window of fixed duration (e.g. 500 kiloyears). Thedynamic window allows for adjustment of the number of data points in the window, so it has aconstant duration in time or space.
Usage
mwin(dat,win,conv=1,verbose=T)
Arguments
dat Your data frame containing stratigraphic data; any number of columns (vari-ables) are permitted, but the first column should be a location identifier (e.g.,depth, height, time).
win Moving window size in units of space or time.
conv Convention for window placement: (1) center each window on a stratigraphiclevel in ’dat’ (DEFAULT), (2) start with the smallest location datum in ’dat’, (3)start with the largest location datum in ’dat’. For options 2 and 3, the center ofthe window will not necessarily coincide with a measured stratigraphic level in’dat’, but edges of the data set are better preserved.
verbose Verbose output? (T or F)
Details
This algorithm steps forward one stratigraphic datum at a time. The output consists of:
Average = this is the average of the depth/time values in the given window.
Center = this is the center of the ’win’ size window.
Midpoint = this is midpoint between first and last observation in the window (for unevenly sampleddata this is typically less than than the size of ’win’).
Value
A data frame containing: Starting index for window, Ending index for window, Location (average),Location (center), Location (midpoint)
mwStats 65
Examples
# generate some noiseex1 <- ar1(npts=50,dt=1)
# jitter sampling timesex1[1]=ex1[1]+rnorm(50,sd=3)# sortex1 = ex1[order(ex1[1],na.last=NA,decreasing=FALSE),]
# run mwinmwin(ex1,win=10)
mwStats ’Dynamic window’ moving average, median and variance of strati-graphic series
Description
’Dynamic window’ moving average, median and variance of stratigraphic series. This routine ad-justs the number of data points in the window so it has a constant duration in time or space, for usewith unevenly sampled data.
Usage
mwStats(dat,cols=NULL,win=NULL,conv=1,output=T,genplot=T,verbose=T)
Arguments
dat Your data frame containing stratigraphic data; any number of columns (vari-ables) are permitted, but the first column should be a location identifier (e.g.,depth, height, time).
cols A vector that identifies the variable column to be extracted (first column auto-matically extracted).
win Moving window size in units of space or time.
conv Convention for window placement: (1) center each window on a stratigraphiclevel in ’dat’ (DEFAULT), (2) start with the smallest location datum in ’dat’, (3)start with the largest location datum in ’dat’. For options 2 and 3, the center ofthe window will not necessarily coincide with a measured stratigraphic level in’dat’, but edges of the data set are better preserved.
output Output results? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Value
A data frame with four columns: Center of window, Average, Median, Variance
66 noKernel
Examples
# generate example series from ar1 noise, 5 kyr sampling intervalex = ar1(npts=1001,dt=5)
# jitter sampling timesex[1]=ex[1]+rnorm(1001,sd=50)# sortex = ex[order(ex[1],na.last=NA,decreasing=FALSE),]
# run mwStatsmwStats(ex,win=100)
noKernel Remove Gaussian kernel smoother from stratigraphic series
Description
Estimate trend and remove from stratigraphic series using a Gaussian kernel smoother
Usage
noKernel(dat,smooth=0.1,sort=F,output=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for smoothing. First column should be location (e.g., depth),second column should be data value.
smooth Degree of smoothing with a Gaussian kernal (0 = no smoothing); for a value of0.5, the kernel is scaled so that its quartiles (viewed as prob densities) are at +/-25 percent of the data series length. Must be > 0.
sort Sort data into increasing depth (required for ksmooth)? (T or F)
output 1= output residual values; 2= output Gaussian kernel smoother.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
bandpass, lowpass, noLow, prewhiteAR, and prewhiteAR1
noLow 67
noLow Fit and remove Lowess smoother from stratigraphic series
Description
Fit and remove lowess smoother from stratigraphic series
Usage
noLow(dat,smooth=.20,output=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for lowess smoother removal. First column should be loca-tion (e.g., depth), second column should be data value.
smooth Lowess smoothing parameter.
output 1= output residual values; 2= output lowess fit
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
bandpass, lowpass, noKernel, prewhiteAR, and prewhiteAR1
pad Pad stratigraphic series with zeros
Description
Pad stratigraphic series with zeros ("zero padding")
Usage
pad(dat,zeros,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for mean removal. First column should be location (e.g.,depth), second column should be data value.
zeros Number of zeros to add on the end of the series. By default, the number of pointswill be doubled.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
68 periodogram
peak Identify maxima of peaks in series, filter at desired threshold value
Description
Identify maxima of peaks in any 1D or 2D series, filter at desired threshold value.
Usage
peak(dat,level,genplot=T,verbose=T)
Arguments
dat 1 or 2 dimensional series. If 2 dimesions, first column should be location (e.g.,depth), second column should be data value.
level Threshold level for filtering peaks. By default all peak maxima reported.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Examples
ex=cycles(genplot=FALSE)peak(ex,level=0.02)
periodogram Simple periodogram
Description
Calculate periodogram for stratigraphic series
Usage
periodogram(dat,padfac=2,demean=T,detrend=F,nrm=1,background=0,output=0,f0=F,fNyq=T,xmin=0,xmax=Nyq,pl=1,genplot=T,verbose=T)
Arguments
dat Stratigraphic series to analyze. First column should be location (e.g., depth),second column should be data value.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints. padfac will automatically promote the total padded series length to aneven number, to ensure the Nyquist frequency is calculated. However, if padfacis set to 0, no padding will be implemented.
periodogram 69
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
nrm Power normalization: 0 = no normalization; 1 = divide Fourier transform bynpts.
background Estimate noise model background spectrum and confidence levels? (0= No, 1=AR1, 2= Power Law)
output Return output as new data frame? (0= no; 1= frequency,amplitude,power,phase(+ background fit and confidence levels, if background selected); 2= frequency,realcoeff.,imag. coeff)
f0 Return results for the zero frequency? (T or F)
fNyq Return results for the Nyquist frequency? (T or F)
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
pl Power spectrum plotting: 1 = log power, 2 = linear power
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
mtm and lowspec
Examples
# ***** PART 1: Demonstrate the impact of tapering# generate example series with 10 periods: 100, 40, 29, 21, 19, 14, 10, 5, 4 and 3 ka.ex=cycles(c(1/100,1/40,1/29,1/21,1/19,1/14,1/10,1/5,1/4,1/3),amp=c(1,.75,0.01,.5,.25,
0.01,0.1,0.05,0.001,0.01))
# set zero padding amount for spectral analyses# (pad= 1 results in no zero padding, pad = 2 will pad the series to two times its original length)# start with pad = 1, then afterwards evaluate pad=2pad=1
# calculate the periodogram with no tapering applied (a "rectangular window")res=periodogram(ex,output=1,padfac=pad)
# save the frequency grid and the power for plottingfreq=res[1]pwr_rect=res[3]
# now compare with results obtained after applying four different tapers:# Hann, 30% cosine taper, DPSS with a time-bandwidth product of 1, and DPSS# with a time-bandwidth product of 3pwr_hann=periodogram(hannTaper(ex,demean=FALSE),output=1,padfac=pad)[3]pwr_cos=periodogram(cosTaper(ex,p=.3,demean=FALSE),output=1,padfac=pad)[3]pwr_dpss1=periodogram(dpssTaper(ex,tbw=1,demean=FALSE),output=1,padfac=pad)[3]pwr_dpss3=periodogram(dpssTaper(ex,tbw=3,demean=FALSE),output=1,padfac=pad)[3]
70 periodogram
# now plot the resultsymin=min(rbind (log(pwr_rect[,1]),log(pwr_hann[,1]),log(pwr_cos[,1]),log(pwr_dpss1[,1]),
log(pwr_dpss3[,1]) ))ymax=max(rbind (log(pwr_rect[,1]),log(pwr_hann[,1]),log(pwr_cos[,1]),log(pwr_dpss1[,1]),
log(pwr_dpss3[,1]) ))
pl(2)plot(freq[,1],log(pwr_rect[,1]),type="l",ylim=c(ymin,ymax),lwd=2,ylab="log(Power)",
xlab="Frequency (cycles/ka)",main="Comparison of rectangle (black), cosine (blue) and Hann (orange) taper",cex.main=1)
lines(freq[,1],log(pwr_hann[,1]),col="orange",lwd=2)lines(freq[,1],log(pwr_cos[,1]),col="blue")points(c(1/100,1/40,1/29,1/21,1/19,1/14,1/10,1/5,1/4,1/3),rep(ymax,10),cex=.5,
col="purple")
plot(freq[,1],log(pwr_rect[,1]),type="l",ylim=c(ymin,ymax),lwd=2,ylab="log(Power)",xlab="Frequency (cycles/ka)",main="Comparison of rectangle (black), 1pi DPSS (green) and 3pi DPSS (red) taper",cex.main=1)
lines(freq[,1],log(pwr_dpss1[,1]),col="green")lines(freq[,1],log(pwr_dpss3[,1]),col="red",lwd=2)points(c(1/100,1/40,1/29,1/21,1/19,1/14,1/10,1/5,1/4,1/3),rep(ymax,10),cex=.5,
col="purple")
# ***** PART 2: Now add a very small amount of red noise to the series# (with lag-1 correlation = 0.5)ex2=exex2[2]=ex2[2]+ar1(rho=.5,dt=1,npts=500,sd=.005,genplot=FALSE)[2]
# compare the original series with the series+noisepl(2)plot(ex,type="l",lwd=2,lty=3,col="black",xlab="time (ka)",ylab="signal",
main="signal (black dotted) and signal+noise (red)"); lines(ex2,col="red")plot(ex[,1],ex2[,2]-ex[,2],xlab="time (ka)",ylab="difference",
main="Difference between the two time series (very small!)")
# calculate the periodogram with no tapering applied (a "rectangular window")res.2=periodogram(ex2,output=1,padfac=pad)
# save the frequency grid and the power for plottingfreq.2=res.2[1]pwr_rect.2=res.2[3]
# now compare with results obtained after applying four different tapers:# Hann, 30% cosine taper, DPSS with a time-bandwidth product of 1, and DPSS# with a time-bandwidth product of 3pwr_hann.2=periodogram(hannTaper(ex2,demean=FALSE),output=1,padfac=pad)[3]pwr_cos.2=periodogram(cosTaper(ex2,p=.3,demean=FALSE),output=1,padfac=pad)[3]pwr_dpss1.2=periodogram(dpssTaper(ex2,tbw=1,demean=FALSE),output=1,padfac=pad)[3]pwr_dpss3.2=periodogram(dpssTaper(ex2,tbw=3,demean=FALSE),output=1,padfac=pad)[3]
pl 71
# now plot the resultsymin=min(rbind (log(pwr_rect.2[,1]),log(pwr_hann.2[,1]),log(pwr_cos.2[,1]),
log(pwr_dpss1.2[,1]),log(pwr_dpss3.2[,1]) ))ymax=max(rbind (log(pwr_rect.2[,1]),log(pwr_hann.2[,1]),log(pwr_cos.2[,1]),
log(pwr_dpss1.2[,1]),log(pwr_dpss3.2[,1]) ))
pl(2)plot(freq.2[,1],log(pwr_rect.2[,1]),type="l",ylim=c(ymin,ymax),lwd=2,ylab="log(Power)",
xlab="Frequency (cycles/ka)",main="Comparison of rectangle (black), cosine (blue) and Hann (orange) taper",cex.main=1)
lines(freq.2[,1],log(pwr_hann.2[,1]),col="orange",lwd=2)lines(freq.2[,1],log(pwr_cos.2[,1]),col="blue")points(c(1/100,1/40,1/29,1/21,1/19,1/14,1/10,1/5,1/4,1/3),rep(ymax,10),cex=.5,
col="purple")
plot(freq.2[,1],log(pwr_rect.2[,1]),type="l",ylim=c(ymin,ymax),lwd=2,ylab="log(Power)",xlab="Frequency (cycles/ka)",main="Comparison of rectangle (black), 1pi DPSS (green) and 3pi DPSS (red) taper",cex.main=1)
lines(freq.2[,1],log(pwr_dpss1.2[,1]),col="green")lines(freq.2[,1],log(pwr_dpss3.2[,1]),col="red",lwd=2)points(c(1/100,1/40,1/29,1/21,1/19,1/14,1/10,1/5,1/4,1/3),rep(ymax,10),cex=.5,
col="purple")
# ***** PART 3: Return to PART 1, but this time increase the zero padding to 2 (pad=2)
pl Set up plots
Description
Open new device and set up for multiple plots, output to screen or PDF if desired.
Usage
pl(n,r,c,h,w,mar,file,title)
Arguments
n Number of plots per page (1-25). When specified, this parameter takes prece-dence and the default settings for r and c are used (the r and c options below areignored).
r Number of rows of plots.
c Number of columns of plots.
h Height of new page (a.k.a. "device").
w Width of new page (a.k.a. "device").
72 plotEha
mar A numerical vector of the form c(bottom, left, top, right) which gives the marginsize specified in inches.
file PDF file name, in quotes. If a file name is not designated, then the plot is outputto the screen instead.
title Plot title (must be in quotes)
plotEha Create color time-frequency plots from eha results.
Description
Create color time-frequency plots from eha results.
Usage
plotEha(spec,xmin,xmax,ymin,ymax,h=6,w=4,ydir=1,pl=0,norm,palette=1,centerZero=T,ncolors=100,colorscale=F,xlab,ylab,filetype=0,output=T,verbose=T)
Arguments
spec Time-frequency spectral results to evaluate. Must have the following format:column 1=frequency; remaining columns (2 to n)=power, amplitude or proba-bility; titles for columns 2 to n must be the location (depth or height). Note thatthis format is ouput by function eha.
xmin Minimum frequency for PLOTTING.
xmax Maximum frequency for PLOTTING.
ymin Minimum depth/height for PLOTTING.
ymax Maximum depth/height for PLOTTING.
h Height of plot in inches.
w Width of plot in inches.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
pl An option for the color plots (0=linear scale; 1=plot log of value [useful forplotting power], 2=normalize to maximum value [useful for plotting amplitude],3=use normalization provided in norm.
norm Optional amplitude normalization divisor, consisting of a single column dataframe.This option is provided in case you’d like to normalize a set of EHA results usingthe same scheme (e.g., before and after removal of spectral lines).
palette What color palette would you like to use? (1) rainbow, (2) grayscale, (3) blue,(4) red, (5) blue-white-red (if values are negative and positive, white is centeredon zero)
centerZero Center color scale on zero (use an equal number of postive and negative colordivisions)? (T or F)
plS 73
ncolors Number of colors steps to use in palette.
colorscale Include a color scale in the plot? (T or F)
xlab Label for x-axis. Default = "Frequency"
ylab Label for y-axis. Default = "Location"
filetype Generate .pdf, .jpeg, .png or tiff file? (0=no; 1=pdf; 2=jpeg; 3=png; 4=tiff)
output If amplitude is normalized (pl = 2), output normalization used? (T or F)
verbose Verbose output? (T or F)
plS Set default plotting parameters for vertical stratigraphic plots
Description
Set default plotting parameters for vertical stratigraphic plots. This is ususally invoked after functionpl.
Usage
plS(f=T,s=1)
Arguments
f Are you plotting the first (leftmost) stratigraphic plot? (T or F)
s Size of the symbols and text on plot. Default = 1
prewhiteAR Prewhiten stratigraphic series with autoregressive filter, order selectedby Akaike Information Criterion
Description
Prewhiten stratigraphic series using autoregressive (AR) filter. Appropriate AR order can be au-tomatically determined using the Akaike Information Criterion, or alternatively, the order may bepredefined.
Usage
prewhiteAR(dat,order=0,method="mle",aic=T,genplot=T,verbose=T)
74 prewhiteAR1
Arguments
dat Stratigraphic series for prewhitening. First column should be location (e.g.,depth), second column should be data value for prewhitening. Series must haveuniform sampling interval.
order AR order for prewhitening (if aic=F), or alternatively, the maximum AR orderto investigate (if aic=T). If order is set to <=0, will evaluate up to maximumdefault order (this varies based on method).
method Method for AR parameter estimation: ("yule-walker", "burg", "ols", "mle","yw")
aic Select model using AIC? if F, will use order. AIC is only strictly valid if methodis "mle".
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
References
Akaike, H. (1974), A new look at the statistical model identification, IEEE Trans. Autom. Control,19, 716-723, doi:10.1109/TAC.1974.1100705.
See Also
ar, arcsinT, bandpass, demean, detrend, divTrend, logT, lowpass, noKernel, and prewhiteAR1
prewhiteAR1 Prewhiten stratigraphic series with AR1 filter, using ’standard’ or un-biased estimate of rho
Description
Prewhiten stratigraphic series using autoregressive-1 (AR1) filter. Rho can be estimated using the’standard’ approach, or following a bias correction.
Usage
prewhiteAR1(dat,setrho=NULL,bias=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for prewhitening. First column should be location (e.g.,depth), second column should be data value for prewhitening. Series must haveuniform sampling interval.
setrho Specified lag-1 correlation coefficient (rho). By default, rho is calculated.
bias Calculate unbiased estimate of rho, as in Mudelsee (2010, eq. 2.45). (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
pwrLaw 75
References
M. Mudelsee, 2010, Climate Time Series Analysis: Classical Statistical and Bootstrap Methods,474 pp., Springer, Dordrecht, Netherlands.
See Also
arcsinT, bandpass, demean, detrend, divTrend, logT, lowpass, noKernel, and prewhiteAR
pwrLaw Generate power law (1/f) noise surrogates
Description
Generate power law (1/f) noise surrogates, following the algorithm of Timmer and Konig (1995).
Usage
pwrLaw(npts=1024,dt=1,mean=0,sdev=1,beta=2,fcut=0,nsim=1,genplot=T,verbose=T)
Arguments
npts number of data points for 1/f surrogate time series
dt sampling interval
mean mean value for 1/f surrogate series
sdev standard deviation for 1/f surrogate series
beta power law coefficient. Positive number will yield a negative slope.
fcut frequency cutoff: below this frequency a plateau will be modeled. Set to zero(default) for no plateau.
nsim Number of surrogate series to generate
genplot generate summary plots (T or F)
verbose verbose output (T or F)
Details
These simulations use the random number generator of Matsumoto and Nishimura (1998). Powerlaw noise series are generated following the algorithm of Timmer and Konig (1995).
References
M. Matsumoto, and T. Nishimura, (1998), Mersenne Twister: A 623-dimensionally equidistributeduniform pseudo-random number generator, ACM Transactions on Modeling and Computer Simu-lation, 8, 3-30.
J. Timmer and K. Konig (1995), On Generating Power Law Noise, Astronomy and Astrophysics:v. 300, p. 707-710.
76 pwrLawFit
pwrLawFit Estimate power law (1/f) fit to power spectrum
Description
Estimate power law (1/f) fit to power spectrum, following the algorithm of Vaughan (2005).
Usage
pwrLawFit(spec,dof=2,flow=NULL,fhigh=NULL,output=1,genplot=T,verbose=T)
Arguments
spec Power spectrum. First column is frequency, second column is raw power (lin-ear). Do not include the zero frequency.
dof Degrees of freedom for power spectral estimate. Default is 2, for a simple peri-odogram.
flow Lowest frequency to include in 1/f fit
fhigh Highest frequency to include in 1/f fit
output Output results of 1/f fit? (0=none; 1=Frequency,Power,Power Law CL,UnbiasedPower Law fit,CL_90,CL_95,CL_99; 2=beta, unbiased log10N, biased log10N)
genplot generate summary plots (T or F)
verbose verbose output (T or F)
References
Vaughan, S. (2005), A simple test for periodic signals in red noise, Astronomy & Astrophysics.
Examples
# generate example series with periods of 400 ka, 100 ka, 40 ka and 20 kaex = cycles(freqs=c(1/400,1/100,1/40,1/20),start=1,end=1000,dt=5)
# add AR1 noisenoise = ar1(npts=200,dt=5,sd=.5)ex[2] = ex[2] + noise[2]
# calculate periodogramres=periodogram(ex,output=1)
pwrLawFit(cb(res,c(1,3)))
rankSeries 77
rankSeries Create lithofacies rank series from bed thickness data
Description
Create lithofacies rank series from bed thickness data.
Usage
rankSeries(dat,dt,genplot=T,verbose=T)
Arguments
dat First column should be bed thickness, and second column should bed lithofaciesrank.
dt Sampling interval for piecewise linear interpolation. By default a grid spacingthat is 5 times smaller than the thinnest bed is used. If dt is set to zero, interpo-lation is skipped.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Examples
# generate example series with random bed thicknessesexThick=rnorm(n=20,mean=10,sd=2)# assign alternating rank of 1 and 2rank=double(20)rank[seq(from=1,to=19,by=2)] <- 1rank[seq(from=2,to=20,by=2)] <- 2
# combine into a dataframeex=cb(exThick,rank)
# generate lithofacies rank seriesrankSeries(ex)
read Read data from file
Description
Read stratigraphic data series from a file, either tab-delimited, CSV, or semicolon-delimited. Firstcolumn MUST contain location data (depth, height, time). The function will remove missing en-tries, sort by location, average duplicate values, and generate summary plots.
78 readMatrix
Usage
read(file=NULL,d=1,h="auto",skip=0,srt=T,ave=T,check=T,genplot=T,verbose=T)
Arguments
file An optional file name, which must be in quotes (use the full directory path if thefile is not in your present working directory). When a file name is not specified(the default), the file will be selected using a graphical user interface.
d What column delimiter is used? (0 = tab/.txt, 1 = comma/.csv, 2 = semicolon).CSV is the default option, which interfaces well with EXCEL.
h Does the data file have column titles/headers? ("yes", "no", "auto"). "auto" willauto detect column titles/headers, which must be single strings and start with acharacter.
skip Number of lines to skip before beginning to read file
srt Sort data values by first column? (T or F)
ave Average duplicate values? (T or F). Only applies if input file has 2 columns
check Check for sorting, duplicates, and empty entries in the data frame? (T or F). Ifset to F, sorting, duplicate averaging and empty entry removal are disabled.
genplot generate summary plots (T or F).
verbose Verbose output? (T or F).
Details
Missing values (in the file that you are reading from) should be indicated by ’NA’. If you haveincluded characters in the column titles that are not permitted by R, they will be modified!
readMatrix Read data matrix from file
Description
Read data matrix from a file, either tab-delimited, CSV, or semicolon-delimited.
Usage
readMatrix(file=NULL,d=1,h="auto",skip=0,output=1,check=T,genplot=F)
Arguments
file An optional file name, which must be in quotes (use the full directory path if thefile is not in your present working directory). When a file name is not specified(the default), the file will be selected using a graphical user interface.
d What column delimiter is used? (0 = tab/.txt, 1 = comma/.csv, 2 = semicolon).CSV is the default option, which interfaces well with EXCEL.
repl0 79
h Does the data file have column titles/headers? ("yes", "no", "auto"). "auto" willauto detect column titles/headers, which must be single strings and start with acharacter.
skip Number of lines to skip before beginning to read file
output Return data as: 1= matrix, 2=data frame
check Check for empty entries in the matrix? (T or F).
genplot generate summary plots (T or F).
Details
Missing values (in the file that you are reading from) should be indicated by ’NA’. If you haveincluded characters in the column titles that are not permitted by R, they will be modified!
repl0 Replace values < 0 with 0
Description
Replace all variable values < 0 with 0. If first column is location ID (depth/height/time), it will notbe processed. Any number of variables (columns) permitted.
Usage
repl0(dat,ID=T,genplot=T,verbose=T)
Arguments
dat Data series to process. If location is included (e.g., depth), it should be in thefirst column.
ID Is a location ID included in the first column? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
80 resample
replEps Replace values <= 0 with smallest positive value
Description
Replace all variable values <= 0 with the smallest positive floating-point number (eps) that can berepresented on machine. If first column is location ID (depth/height/time), it will not be processed.Any number of variables (columns) permitted.
Usage
replEps(dat,ID=T,genplot=T,verbose=T)
Arguments
dat Data series to process. If location is included (e.g., depth), it should be in thefirst column.
ID Is a location ID included in the first column? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
resample Resample stratigraphic series
Description
Resample a stratigraphic series using a new (variably sampled) time or space axis. Values arepiecewise-linearly interpolated from original data.
Usage
resample(dat,xout,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for resampling. First column should be location (e.g., depth),second column should be data value.
xout Vector of new sampling locations.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
rmNA 81
rmNA Remove stratigraphic levels that contain one or more NAs
Description
Remove stratigraphic levels that contain one or more NAs.
Usage
rmNA(dat,genplot=T,verbose=T)
Arguments
dat Data series to process. If location is included (e.g., depth), it should be in thefirst column.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
s Standardize variable in stratigraphic series
Description
Standardize variable in stratigraphic series (subtract mean value and divide by standard deviation)
Usage
s(dat,genplot=F,verbose=T)
Arguments
dat Stratigraphic series for standardization. First column should be location (e.g.,depth), second column should be data value.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
82 sedrate2time
sedRamp Apply ’ramping’ sedimentation rate model to convert time to stratig-raphy
Description
Apply a linearly increasing (or decreasing) sedimentation rate model to convert time to stratigraphy.
Usage
sedRamp(dat,srstart=0.01,srend=0.05,genplot=T,verbose=T)
Arguments
dat Time series. First column should be time (in ka), second column should be datavalue.
srstart Initial sedimentation rate (in m/ka).
srend Final sedimentation rate (in m/ka).
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Value
modeled stratigraphic series.
Examples
# generate example series with 3 precession terms using function 'cycles'# then convert from time to space using sedimentation rate that increases from 1 to 7 cm/kaex=sedRamp(cycles(),srstart=0.01,srend=0.07)
sedrate2time Integrate sedimentation rate curve to obtain time-space map
Description
Integrate sedimentation rate curve to obtain time-space map.
Usage
sedrate2time(sedrates,timedir=1,genplot=T,check=T,verbose=T)
slideCor 83
Arguments
sedrates Data frame containing depth/height in first column (meters) and sedimentationrates in second column (cm/ka).
timedir Floating time scale direction: 1= time increases with depth/height; 2= time de-creases with depth/height.)
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (T or F)
slideCor Identify optimal spatial/temporal shift to maximize correlation be-tween two stratigraphic series.
Description
Identify optimal spatial/temporal shift to maximize correlation between two stratigraphic series.
Usage
slideCor(dat1,dat2,rev=F,cormethod=1,minpts=5,output=T,genplot=T,verbose=T)
Arguments
dat1 Stratigraphic series 1. First column should be location (e.g., depth), secondcolumn should be data value.
dat2 Stratigraphic series 2. First column should be location (e.g., depth), secondcolumn should be data value.
rev Reverse polarity of stratigraphic series 2 (multiply proxy data value by -1)? (Tor F)
cormethod Method used for calculation of correlation coefficient (1=Pearson, 2=Spearmanrank, 3=Kendall)
minpts Minimum number of data points for calculation of correlation coefficient.
output Output correlation coefficient results as a dataframe? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
84 slideCor
Details
slideCor is a general purpose tool to identify the optimal spatial/temporal correlation between twodata sets. A few example applications include: (1) stratigraphic correlation of data series fromtwo locations (as in Preto et al., 2004), (2) identification of the optimal spatial/temporal lag be-tween two variables from the same site, and (3) identification of the optimal fit between a floatingastrochronology and astronomical target (e.g, Mitchell et al., 2008).
Both series must be evenly sampled, but are not required to have the same sampling interval. Ifstratigraphic series of different duration/length are being compared, the shift (in spatial or temporalunits) should be interpreted as the location within the longer stratigraphic series where the shorterstratigraphic series begins. If both stratigraphic series are of the same duration/length, then the shiftis the location within dat1 where dat2 begins.
In some cases, it may be desirable to smooth or bandpass the data series before implementingslideCor (e.g., functions noLow, noKernel, bandpass, taner, etc.).
References
Preto, N., Hinnov, L.A., De Zanche, V., Mietto, P., and Hardie, L.A., 2004, The Milankovitchinterpretation of the Latemar Platform cycles (Dolomites, Italy): Implications for geochronology,biostratigraphy, and Middle Triassic carbonate accumulation, SEPM Special Publication 81.
Mitchell, R.N., Bice, D.M., Montanari, A., Cleaveland, L.C., Christianson, K.T., Coccioni, R., andHinnov, L.A., 2008, Oceanic anoxic cycles? Orbital prelude to the Bonarelli Level (OAE 2), EarthPlanet. Sci. Lett. 26, 1-16.
See Also
surrogateCor
Examples
# Example 1: generate AR1 noiseex1 <- ar1(npts=1000,dt=1)# isolate a sectionex2 <- iso(ex1,xmin=200,500)ex2[1] <- ex2[1]-200
slideCor(ex1,ex2)
# Example 2: an astronomical signalex1=etp(tmin=0,tmax=1000)# isolate a sectionex2=iso(ex1,xmin=400,xmax=600)ex2[1] <- ex2[1]-400
slideCor(ex1,ex2)
sortNave 85
sortNave Remove missing entries, sort data, average duplicates
Description
Sort and average duplicates in stratigraphic series, as performed in ’read’ function.
Usage
sortNave(dat,sortDecr=F,ave=T,xmin=NULL,xmax=NULL,genplot=1,verbose=T)
Arguments
dat Stratigraphic series for processing. First column should be location (e.g., depth),second column should be data value.
sortDecr Sorting direction? (F=increasing, T=decreasing)
ave Average duplicate values? (T or F)
xmin Minimum x-axis value for plotting
xmax Maximum x-axis value for plotting
genplot Generate summary plots? 0=none, 1=stratigraphic series, distribution, box plot,Q-Q, 2=stratigraphic series
verbose Verbose output? (T or F)
stepHeat Ar/Ar Geochronology: Generate an Ar/Ar age spectrum and calculatestep-heating plateau age.
Description
The stepHeat function will evaluate data from stepwise heating experiments, producing an Ar/Arage spectrum, a weighted mean age with uncertainty, and other helpful statistics/plots (with inter-active graphics for data culling). The function includes the option to generate results using theapproach of IsoPlot 3.70 (Ludwig, 2008) or ArArCALC (Koppers, 2002).
Usage
stepHeat(dat,unc=1,lambda=5.463e-10,J=NULL,Jsd=NULL,CI=2,cull=-1,del=NULL,output=F,idPts=T,size=NULL,unit=1,setAr=95,color="black",genplot=T,verbose=T)
86 stepHeat
Arguments
dat dat must be a data frame with seven columns, as follows: (1) %Ar39 released,(2) date, (3) date uncertainty (one or two sigma), (4) K/Ca, (5) %Ar40*, (6) F,and (7) F uncertainty (one or two sigma). NOTE: F is the ratio Ar40*/Ar39K(see Koppers, 2002).
unc What is the uncertainty on your input dates? (1) one sigma, or (2) two sigma.DEFAULT is one sigma. This also applies to the F uncertainty, and the J-valueuncertainty (if specified)
lambda Total decay constant of K40, in units of 1/year. The default value is 5.463e-10/year (Min et al., 2000).
J Neutron fluence parameter
Jsd Uncertainty for J-value (neutron fluence parameter; as one or two sigma)
CI Which convention would you like to use for the 95% confidence intervals? (1)ISOPLOT (Ludwig, 2008), (2) ArArCALC (Koppers, 2002)
cull Would you like select dates with a graphical interface? (0=no, 1=select pointsto retain, -1=select points to remove)
del A vector of indices indicating dates to remove from weighted mean calculation.If specified, this takes precedence over cull.
output Return weighted mean results as new data frame? (T or F)
idPts Identify datum number on each point? (T or F)
size Multiplicative factor to increase or decrease size of symbols and fonts. Thedefault is 1.4
unit The time unit for your results. (1) = Ma, (2) = Ka
setAr Set the %Ar40* level to be illustrated on the plot. The default is 95%.
color Color to use for symbols. Default is black.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
This function performs weighted mean age calculations for step-heating data, including estimationof age uncertainties, mean square weighted deviation, and probability of fit.
The following plots are produced:
(1) %Ar40* versus %Ar39 released
(2) K/Ca versus %Ar39 released
(3) Ar/Ar age spectrum, with 2 sigma uncertainties for each step, and weighted mean with 95%confidence interval (in red)
If the J-value and its uncertainty are input, stepHeat will calculate and include the uncertaintyassociated with J. The uncertainty is calculated and propagated following equation 18 of Koppers(2002).
A NOTE regarding confidence intervals: There are two conventions that can be used to calculatethe confidence intervals, selected with the option ’CI’:
stepHeat 87
(1) ISOPLOT convention (Ludwig, 2008). When the probability of fit is >= 0.15, the confidenceinterval is based on 1.96*sigma. When the probability of fit is < 0.15, the confidence interval isbased on t*sigma*sqrt(MSWD).
(2) ArArCALC convention (Koppers, 2002). When MSWD <=1, the confidence interval is basedon 1.96*sigma. When MSWD > 1, the confidence interval is based on 1.96*sigma*sqrt(MSWD).
ADDITIONAL ADVICE: Use the function readMatrix to load your data in R (rather than the func-tion read).
References
A.A.P. Koppers, 2002, ArArCALC- software for 40Ar/39Ar age calculations: Computers & Geo-sciences, v. 28, p. 605-619.
K.R. Ludwig, 2008, User’s Manual for Isoplot 3.70: A Geochronological Toolkit for MicrosoftExcel: Berkeley Geochronology Center Special Publication No. 4, Berkeley, 77 p.
I. McDougall and T.M. Harrison, 1991, Geochronology and Thermochronology by the 40Ar/39ArMethod: Oxford University Press, New York, 269 pp.
K. Min, R. Mundil, P.R. Renne, and K. Ludwig, 2000, A test for systematic errors in 40Ar/39Argeochronology through comparison with U/Pb analysis of a 1.1-Ga rhyolite: Geochimica et Cos-mochimica Acta, v. 64, p. 73-98.
I. Wendt and C. Carl, 1991, The statistical distribution of the mean squared weighted deviation:Chemical Geology, v. 86, p. 275-285.
See Also
wtMean
Examples
# Check to see if this is an interactive R session, for compliance with CRAN standards.# YOU CAN SKIP THE FOLLOWING LINE IF YOU ARE USING AN INTERACTIVE SESSION.if(interactive()) {
# Sample MT-09-09 incremental heating Ar/Ar data from Sageman et al. (2014).perAr39 <- c(4.96,27.58,19.68,39.9,6.25,1.02,0.42,0.19)age <- c(90.08,89.77,89.92,89.95,89.89,89.55,87.71,86.13)sd <- c(0.18,0.11,0.08,0.06,0.14,0.64,1.5,3.22)KCa <- c(113,138,101,195,307,27,17,24)perAr40 <- c(93.42,99.42,99.64,99.79,99.61,97.99,94.64,90.35)Fval <- c(2.148234,2.140643,2.144197,2.145006,2.143627,2.135163,2.090196,2.051682)Fsd <- c(0.00439,0.00270,0.00192,0.00149,0.00331,0.01557,0.03664,0.07846)ex <- data.frame(cbind(perAr39,age,sd,KCa,perAr40,Fval,Fsd))
stepHeat(ex)
# plot without points identifiedstepHeat(ex,size=0,idPts=FALSE,cull=0)
}
88 surrogateCor
strats Summary statistics for stratigraphic series
Description
Summary statistics for stratigraphic series: sampling interval and proxy values.
Usage
strats(dat,output=0,genplot=1)
Arguments
dat Stratigraphic series to evaluate. The first column should contain location (e.g.,depth), and the second column should contain data value. This function alsoaccepts non-stratigraphic (single column) input, in which case the sampling in-terval assessment is skipped.
output Output: (0) nothing, (1) cumulative dt as percent of data points, (2) cumulativedt as percent of total interval duration, (3) dt by location
genplot Generate summary plots? (0) none, (1) include plot of cumulative dt, (2) includedt histogram/density plot
Details
This function will generate a range of summary statistics for time series, including sampling intervalinformation and the statistical distribution of proxy values.
surrogateCor Estimate correlation coefficient and significance for serially corre-lated data
Description
Estimate correlation coefficient and significance for serially correlated data. This algorithm permitsthe analysis of data sets with different sampling grids, as discussed in Baddouh et al. (2016). Thesampling grid from the data set with fewer points (in the common interval) is used for resampling.Resampling is conducted using piecewise-linear interpolation.
If either dat1 or dat2 have only one column, the resampling is skipped.
The significance of the correlation is determined using the method of Ebisuzaki W. (1997).
Usage
surrogateCor(dat1,dat2,firstDiff=F,cormethod=1,nsim=1000,output=2,genplot=T,verbose=T)
surrogateCor 89
Arguments
dat1 Data series with one or two columns. If two columns, first should be location(e.g., depth), second column should be data value.
dat2 Data series with one or two columns. If two columns, first should be location(e.g., depth), second column should be data value.
firstDiff Calculate correlation using first differences? (T or F)
cormethod Method used for calculation of correlation coefficient (1=Pearson, 2=Spearmanrank, 3=Kendall)
nsim Number of phase-randomized surrogate series to generate. If nsim <=1, simula-tion is deactivated.
output Return which of the following?: 1= correlation coefficients for each simulation;2= correlation coefficient for data series; 3= data values used in correlation esti-mate (resampled)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
Paraphrased from Baddouh et al. (2016): To provide a quantitative evaluation of the correlationbetween two data sets that do not share a common sampling grid, we introduce a statistical approachthat employs sample interpolation, and significance testing with phase-randomized surrogate data(Ebisuzaki, 1997). The sparser sampling grid is used to avoid over-interpolation. Correlation isevaluated using Pearson, Spearman Rank, or Kendall rank coefficients. The statistical significanceof the resulting correlation coefficients are estimated via Monte Carlo simulations using phase-randomized surrogates; the surrogates are subject to the same interpolation process, and compensatefor potential serial correlation of data (Ebisuzaki, 1997).
The first-difference series of each variable can also evaluated, to assess correlation in the magnitudeof change between sequential stratigraphic samples rather than absolute magnitude.
References
M. Baddouh, S.R. Meyers, A.R. Carroll, B.L. Beard, C.M. Johnson , 2016, Lacustrine 87-Sr/86-Sras a tracer to reconstruct Milankovitch forcing of the Eocene hydrologic cycle: Earth and PlanetaryScience Letters.
W. Ebisuzaki, 1997, A Method to Estimate the Statistical Significance of a Correlation When theData Are Serially Correlated: Journal of Climate, v. 10, p. 2147-2153.
See Also
surrogates
Examples
# generate two stochastic AR1 seriesex1 <- ar1(npts=100,dt=5)ex2 <- ar1(npts=100,dt=6)
90 surrogates
# calculate pearson correlation coefficient and p-valuesurrogateCor(ex1,ex2)
surrogates Generate phase-randomized surrogate series as in Ebisuzaki (1997)
Description
Generate phase-randomized surrogate series as in Ebisuzaki (1997).
Usage
surrogates(dat,nsim=1,preserveMean=T,std=T,genplot=T,verbose=T)
Arguments
dat Data series with one or two columns. If two columns, first should be location(e.g., depth), second column should be data value.
nsim Number of phase-randomized surrogate series to generate.
preserveMean Should surrogate series have the same mean value as data series? (T or F)
std Standardize results to guarantee equivalent variance as data series? (T or F)
genplot Generate summary plots? Only applies if nsim=1. (T or F)
verbose Verbose output? (T or F)
Details
This function will generate phase-randomized surrogate series as in Ebisuzaki (1997). It is an R-translation of the Matlab code by V. Moron (see link below), with modifications and additionalfeatures.
References
W. Ebisuzaki, 1997, A Method to Estimate the Statistical Significance of a Correlation When theData Are Serially Correlated: Journal of Climate, v. 10, p. 2147-2153.
Matlab code by V. Moron: http://www.mathworks.com/matlabcentral/fileexchange/10881-weaclim/content/ebisuzaki.m
Original C-code by W. Ebisuzaki: http://www.ftp.cpc.ncep.noaa.gov/wd51we/random_phase/
synthStrat 91
Examples
# generate example series with 3 precession terms and noiseex <- cycles(start=0,end=500,noisevar=.0004,dt=5)
# generate phase-randomized surrogatesran_ex <- surrogates(ex,nsim=1)
# compare periodograms of data and surrogatesres1 <- periodogram(ex,padfac=0,output=1,genplot=FALSE)res2 <- periodogram(ran_ex,padfac=0,output=1,genplot=FALSE)
pl(2)plot(ex,type="l",main="black=original; red=surrogate")lines(ran_ex,col="red",lty=4)plot(res1[,1],res1[,2],type="l",lwd=2,main="black=original; red=surrogate",
xlab="frequency",ylab="amplitude")lines(res2[,1],res2[,2],col="red",lwd=2,lty=4)
synthStrat Synthesize stratigraphy from forcing function
Description
Synthesize stratigraphy from forcing function.
Usage
synthStrat(signal=NULL,nfacies=4,clip=T,flip=F,fmax=0.1,output=F,genplot=2,verbose=T)
Arguments
signal Forcing signal. First column should be time (in ka), second column should beforcing.
nfacies Number of sedimentary facies to model.
clip Clip forcing signal at mean value? (T or F)
flip Reverse the sign of the forcing? (T or F)
fmax Maximum frequency for spectra (if genplot=2).
output Output facies series? (T or F)
genplot Generate summary plots? (1) plot stratigraphy, (2) plot statigraphy and spectra.
verbose Verbose output? (T or F)
Value
modeled stratigraphic series.
92 taner
Examples
# EX.1: precession, unclippedsignal=etp(tmin=8400,tmax=8900,pWt=1,oWt=0,eWt=0)synthStrat(signal,nfacies=4,clip=FALSE,genplot=2)
# EX.2: more finely resolved facies#synthStrat(signal,nfacies=15,clip=FALSE,genplot=2)
# EX.3: couplets#synthStrat(signal,nfacies=2,clip=FALSE,genplot=2)
# EX.4: precession, clipped#synthStrat(signal,nfacies=4,genplot=2)
# EX.5: noisenoise=ar1(npts=501,rho=0.8)#synthStrat(noise,nfacies=4,genplot=2)
# EX.6: precession + noise#signal2=signal#signal2[2]=signal2[2]+0.75*noise[2]#synthStrat(signal2,nfacies=4,genplot=2)
# EX.7: p-0.5t, clipped (demonstrates interference pattern; compare with EX.4#signal3=etp(tmin=8400,tmax=8900,pWt=1,oWt=-0.5,eWt=0)#synthStrat(signal3,nfacies=4,genplot=2)
# EX.8: ice sheet model, using p-0.5t#ice=imbrie()#synthStrat(ice,nfacies=5,clip=FALSE,genplot=2)
# EX.9: precession, clipped, ramping sedimentation rate#synthStrat(linterp(sedRamp(signal,genplot=FALSE),genplot=FALSE),nfacies=6,# clip=TRUE,genplot=2,fmax=10)
taner Apply Taner bandpass or lowpass filter to stratigraphic series
Description
Apply Taner bandpass or lowpass filter to stratigraphic series. This function can also be used tonotch filter or highpass a record (see examples).
Usage
taner(dat,padfac=2,flow=NULL,fhigh=NULL,roll=10^3,demean=T,detrend=F,addmean=T,output=1,xmin=0,xmax=Nyq,genplot=T,check=T,verbose=T)
taner 93
Arguments
dat Stratigraphic series for bandpass filtering. First column should be location (e.g.,depth), second column should be data value.
padfac Pad with zeros to (padfac*npts) points, where npts is the original number of datapoints.
flow Low frequency cut-off for Taner filter (half power point). If this value is not set(NULL), it will default to -1*fhigh, which will create a lowpass filter.
fhigh High frequency cut-off for Taner filter (half power point).
roll Roll-off rate, in dB/octave. Typical values are 10^3 to 10^12, but can be larger.
demean Remove mean from data series? (T or F)
detrend Remove linear trend from data series? (T or F)
addmean Add mean value to bandpass result? (T or F)
output Output: (1) filtered series, (2) bandpass filter window.
xmin Smallest frequency for plotting.
xmax Largest frequency for plotting.
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (T or F)
Value
bandpassed stratigraphic series.
References
http://www.rocksolidimages.com/pdf/attrib_revisited.htm#_Toc328470897
See Also
bandpass, lowpass, noKernel, noLow, prewhiteAR, and prewhiteAR1
Examples
# generate example series with periods of 405 ka, 100 ka, 40ka, and 20 ka, plus noiseex=cycles(freqs=c(1/405,1/100,1/40,1/20),end=1000,dt=5,noisevar=.1)
# bandpass precession term using Taner windowbandpass_ex <- taner(ex,flow=0.045,fhigh=0.055,roll=10^10)
# lowpass filter eccentricity terms using Taner windowlowpass_ex=taner(ex,fhigh=.02,roll=10^10)
# notch filter (remove) obliquity term using Taner window# if you'd like the final notch filtered record to be centered on the mean proxy# value, set addmean=FALSE
94 testBackground
notch_ex <- taner(ex,flow=0.02,fhigh=0.03,roll=10^10,addmean=FALSE)notch_ex[2] <- ex[2]-notch_ex[2]pl(2)plot(ex,type="l",main="Eccentricity+Obliquity+Precession")plot(notch_ex,type="l",main="Following application of obliquity notch filter")
# highpass filter obliquity and precession terms using Taner window# if you'd like the final highpass filtered record to be centered on the mean proxy# value, set addmean=FALSEhighpass_ex=taner(ex,fhigh=.02,roll=10^10,addmean=FALSE)highpass_ex[2] <- ex[2]-highpass_ex[2]pl(2)plot(ex,type="l",main="Eccentricity+Obliquity+Precession")plot(highpass_ex,type="l",main="Obliquity+Precession highpassed signal")
testBackground Evaluate power spectrum false positive rates via Monte Carlo simula-tion
Description
This is a simulation tool to evaluate power spectrum false positive rates, the frequency distributionof the false positives, and the behavior of numerous "multiple correction" procedures, for a rangeof background estimation approaches that are implemented in Astrochron. The tool can be usedto conduct surrogate analyses, alongside analysis of real data, to better understand the suitabilityof particular background estimation approaches. The resulting simulations are similar to thosepresented in Figure 3 of Meyers (2012) and Crampton et al. (PNAS).
Usage
testBackground(npts=1001,dt=5,noiseType="ar1",coeff=NULL,method="periodogramAR1",opt=NULL,demean=T,detrend=F,low=0,tbw=3,multi=F,iter=2000,output=F,
verbose=T,genplot=F)
Arguments
npts Number of points in simulated stratigraphic series (surrogates).
dt Sampling interval for surrogates.
noiseType Select "ar1" for AR1 noise surrogates, or "pwrLaw" for Power Law noise surro-gates
coeff AR1 coefficient (rho) or Power Law coefficient (beta) for surrogates.
method Background estimation method: (1) "mtmAR1" (function mtm), (2) "mtmML96"(function mtmML96), (3) "lowspec" (function lowspec), (4) "mtmPL" (functionmtmPL), (5) "periodogramPL" (function periodogram), (6) "periodogramAR1"(function periodogram)
opt Method specific options. For mtmML96, this is medsmooth (see function mt-mML96); for lowspec this is lowspan (see function lowspec); for periodogramthis is percent cosine taper (see function cosTaper).
testBackground 95
demean Remove mean value from simulated surrogates? (T or F; this option does notapply to lowspec)
detrend Remove linear trend from simulated surrogates? (T or F)low Remove long-term trend using a LOWESS smoother? Choose a value ranging
from 0-1 (see function noLow). 0 = no long-term trend removal.tbw MTM time-bandwidth product. This option is ignored for methods 5 and 6.multi Evaluate a range of multiple-comparison tests too? (T or F)iter Number of iterations (surrogate series) for Monte Carlo simulation.output Output data frame? (T or F)verbose Verbose output? (T or F)genplot Generate summary plots? (T or F)
Details
The Monte Carlo simulations can utilize AR1 or Power Law noise surrogates. Background estima-tion approaches include conventional AR1, ML96, LOWSPEC and Power Law. The function alsoallows evaluation of common data detrending approaches (linear trend removal, LOWESS trendremoval).
Note that MTM-ML96 conducts the Mann and Lees (1996; ML96) "robust red noise" analysis, withan improved median smoothing approach. The original Mann and Lees (1996) approach applies atruncation of the median smoothing window to include fewer frequencies near the edges of thespectrum; while truncation is required, its implementation in the original method often results inan "edge effect" that can produce excess false positive rates at low frequencies, commonly withinthe eccentricity-band (Meyers, 2012). To help address this issue, an alternative median smoothingapproach is applied that implements Tukey’s robust end-point rule and symmetrical medians (see thefunction mtmML96 for more details). This version of the ML96 algorithm was first implementedin Patterson et al. (2014).
See function multiTest for more information on the multiple comparison tests evaluated.
References
W.S. Cleveland, 1979, Locally weighted regression and smoothing scatterplots: Journal of theAmerican Statistical Association, v. 74, p. 829-836.
J.S. Campton, S.R. Meyers, R.A. Cooper, P.M Sadler, M. Foote, D. Harte, 2018, Pacing of Paleozoicmacroevolutionary rates by Milankovitch grand cycles: Proceedings of the National Academy ofSciences, doi:10.1073/pnas.1714342115.
M.E. Mann, and J.M. Lees, 1996, Robust estimation of background noise and signal detection inclimatic time series, Clim. Change, 33, 409-445.
S.R. Meyers, 2012, Seeing Red in Cyclic Stratigraphy: Spectral Noise Estimation for Astrochronol-ogy: Paleoceanography, 27, PA3228, doi:10.1029/2012PA002307.
M.O. Patterson, R. McKay, T. Naish, C. Escutia, F.J. Jimenez-Espejo, M.E. Raymo, M.E., S.R.Meyers, L. Tauxe, H. Brinkhuis, and IODP Expedition 318 Scientists,2014, Response of the EastAntarctic Ice Sheet to orbital forcing during the Pliocene and Early Pleistocene, Nature Geoscience,v. 7, p. 841-847.
D.J. Thomson, 1982, Spectrum estimation and harmonic analysis: IEEE Proceedings, v. 70, p.1055-1096.
96 testPrecession
See Also
confAdjust,multiTest,lowspec, mtm, mtmML96, mtmPL, and periodogram
Examples
# evaluate false positive rate for MTM-AR1 using AR1 surrogatestestBackground(noiseType="ar1",method="mtmAR1")
# evaluate false positive rate for MTM-AR1 using Power Law surrogatestestBackground(noiseType="pwrLaw",method="mtmAR1")
testPrecession Astrochronologic testing via the precession amplitude modulation ap-proach of Zeeden et al. (2015).
Description
Astrochronologic testing via the precession amplitude modulation approach of Zeeden et al. (2015),as updated in Zeeden et al. (2018 submitted).
Usage
testPrecession(dat,nsim=1000,gen=1,edge=0.025,maxNoise=1,rho=NULL,detrendEnv=T,solution=NULL,output=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series to analyze. First column should be location (time in ka, apositive value), second column should be data value.
nsim Number of Monte Carlo simulations (phase-randomized surrogates or AR1 sur-rogates).
gen Monte Carlo simulation generator: (1) use phase-randomized surrogates, (2) useAR1 surrogates.
edge Percentage of record to exclude from beginning and end of data series, to removeedge effects. (a value from 0-1)
maxNoise Maximum noise level to add in simulations. A value of 1 will apply maximumnoise that is equivalent to 1 standard deviation of the data.
rho Specified lag-1 correlation coefficient (rho). If rho is not specified, it will becalculated within the function.
detrendEnv Linearly detrend envelope? (T or F)
solution Theoretical solution used for astrochronologic testing. Solution should be inthe format: time (ka), precession angle, obliquity, eccentricity (the output fromfunction ’getLaskar’). By default this is automatically determined within thefunction, using the solution of Laskar et al. (2004).
testPrecession 97
output Return results as a new data frame? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
This astrochronologic testing method compares observed precession-scale amplitude modulationsto those expected from the theoretical eccentricity solutions. It is applicable for testing astrochronolo-gies spanning 0-50 Ma. The technique implements a series of filters to guard against artificial intro-duction of eccentricity modulations during tuning and data processing, and evaluates the statisticalsignificance of the results using Monte Carlo simulation (Zeeden et al., 2015).
The algorithm includes an improvement in the significance testing approach. Specifically, as asafeguard against artificially imposed modulations, an adaptive noise addition step is implemented(as outlined in Zeeden et al., submitted).
The astronomically-tuned data series under evaluation should consist of two columns: time in kilo-years & data value. Note that time must be positive. The default astronomical solutions used for theastrochronologic testing come from Laskar et al. (2004).
When reporting a p-value for your result, it is important to consider the number of simulations used.A factor of 10 is appropriate, such that for 1000 simulations one would report a minimum p-valueof "p<0.01", and for 10000 simulations one would report a minimum p-value of "p<0.001".
Please be aware that the kernel density estimate plots, which summarize the simulations, represent’smoothed’ models. Due to the smoothing bandwidth, they can sometimes give the impression ofsimulation values that are larger or smaller than actually present. However, the reported p-valuedoes not suffer from these issues.
IMPORTANT CHANGES (June 20, 2018): Note that this version has been updated to use ’solution’instead of ’esinw’, for consistency with the function ’testTilt’. If you are invoking the default option,you do not need to make any changes to your script. Also note that the new option ’edge’ has beenadded, which by default will truncate your data series by 5 percent (2.5 percent on each end of therecord), to guard against edge effects that can be present in the amplitude envelope. Set edge to 0to reconstruct the original (now legacy) ’testPrecession’ approach.
Value
When nsim is set to zero, the function will output a data frame with five columns:
1=time, 2=precession bandpass filter output, 3=amplitude envelope of (2), 4=lowpass filter out-put of (3), 5=theoretical eccentricity (as extracted from precession modulations using the filteringalgorithm), 6=(2) + noise, 7=amplitude envelope of (6), 8=lowpass filter output of (7)
When nsim is > 0, the function will output the correlation coefficients for each simulation.
References
C. Zeeden, S.R. Meyers, L.J. Lourens, and F.J. Hilgen, 2015, Testing astronomically tuned agemodels: Paleoceanography, 30, doi:10.1002/2014PA002762.
C. Zeeden, S.R. Meyers, F.J. Hilgen, L.J. Lourens, and J. Laskar, submitted, Time scale evaluationand the quantification of obliquity forcing: Quaternary Science Reviews.
98 testTilt
J. Laskar, P. Robutel, F. Joutel, M. Gastineau, A.C.M. Correia, and B. Levrard, B., 2004, A longterm numerical solution for the insolation quantities of the Earth: Astron. Astrophys., Volume 428,261-285.
See Also
asm, eAsmTrack, timeOpt, and timeOptSim
Examples
### as a test series, use the three dominant precession terms from Berger et al. (1992)ex<-cycles(start=0,end=1000,dt=2)
### now conduct astrochronologic testingres1=testPrecession(ex)
### if you plan to run testPrecession repeatedly, it is advisable to download the astronomical### solution firstsolution<-getLaskar()
### now conduct astrochronologic testingres2<-testPrecession(ex,solution=solution)
testTilt Astrochronologic testing via the obliquity amplitude modulation ap-proach of Zeeden et al. (2019 submitted).
Description
Astrochronologic testing via the obliquity amplitude modulation approach of Zeeden et al. (2019submitted).
Usage
testTilt(dat,nsim=1000,gen=1,edge=0.025,cutoff=1/150,maxNoise=0.25,rho=NULL,detrendEnv=T,solution=NULL,output=F,genplot=T,verbose=T)
Arguments
dat Stratigraphic series to analyze. First column should be location (time in ka, apositive value), second column should be data value.
nsim Number of Monte Carlo simulations (phase-randomized surrogates or AR1 sur-rogates).
gen Monte Carlo simulation generator: (1) use phase-randomized surrogates, (2) useAR1 surrogates.
testTilt 99
edge Percentage of record to exclude from beginning and end of data series, to removeedge effects. (0-1)
cutoff Cutoff frequency for lowpass filtering.
maxNoise Maximum noise level to add in simulations. A value of 1 will apply maximumnoise that is equivalent to 1 sd of data.
rho Specified lag-1 correlation coefficient (rho). This value is only used if gen=2. Ifrho is not specified, it will be calculated within the function.
detrendEnv Linearly detrend envelope? (T or F)
solution Theoretical solution used for astrochronologic testing. Solution should be inthe format: time (ka), precession angle, obliquity, eccentricity (the output fromfunction ’getLaskar’). By default this is automatically determined within thefunction, using the solution of Laskar et al. (2004).
output Return results as a new data frame? (T or F)
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
This astrochronologic testing method compares observed obliquity-scale amplitude modulations tothose expected from the theoretical solutions. It is applicable for testing astrochronologies spanning0-50 Ma. The technique implements a series of filters to guard against artificial introduction ofmodulations during tuning and data processing, and evaluates the statistical significance of theresults using Monte Carlo simulation. The algorithm includes an adaptive noise addition step toimprovement the significance testing approach. See Zeeden et al. (2019 submitted) for additionalinformation.
The astronomically-tuned data series under evaluation should consist of two columns: time in kilo-years & data value. Note that time must be positive. The default obliquity solution used for theastrochronologic testing comes from Laskar et al. (2004).
When reporting a p-value for your result, it is important to consider the number of simulations used.A factor of 10 is appropriate, such that for 1000 simulations one would report a minimum p-valueof "p<0.01", and for 10000 simulations one would report a minimum p-value of "p<0.001".
Please be aware that the kernel density estimate plots, which summarize the simulations, represent’smoothed’ models. Due to the smoothing bandwidth, they can sometimes give the impression ofsimulation values that are larger or smaller than actually present. However, the reported p-valuedoes not suffer from these issues.
Value
When nsim is set to zero, the function will output a data frame with five columns:
1=time, 2=obliquity bandpass filter output, 3=amplitude envelope of (2), 4=lowpass filter output of(3), 5=theoretical obliquity (as extracted from modulations using the filtering algorithm), 6=(2) +noise, 7=amplitude envelope of (6), 8=lowpass filter output of (7)
When nsim is > 0, the function will output the correlation coefficients for each simulation.
100 timeOpt
References
C. Zeeden, S.R. Meyers, F.J. Hilgen, L.J. Lourens, and J. Laskar, 2019 submitted, Time scale eval-uation and the quantification of obliquity forcing: Quaternary Science Reviews.
C. Zeeden, S.R. Meyers, L.J. Lourens, and F.J. Hilgen, 2015, Testing astronomically tuned agemodels: Paleoceanography, 30, doi:10.1002/2014PA002762.
J. Laskar, P. Robutel, F. Joutel, M. Gastineau, A.C.M. Correia, and B. Levrard, B., 2004, A longterm numerical solution for the insolation quantities of the Earth: Astron. Astrophys., Volume 428,261-285.
See Also
asm, eAsmTrack, timeOpt, and timeOptSim
Examples
### as a test series, use the obliquity series from Laskar et al. (2004), spanning### the past 4 million yearsex<-etp(tmin=0,tmax=4000,dt=2,eWt=0,oWt=1,pWt=0,solution=solution,standardize=FALSE)
### now conduct astrochronologic testingres1=testTilt(ex)
### if you plan to run testTilt repeatedly, it is advisable to download the astronomical### solutionsolution<-getLaskar()
### now conduct astrochronologic testingres<-testTilt(ex,solution=solution)
timeOpt TimeOpt: Evaluation of eccentricity-related amplitude modulationand bundling in paleoclimate data
Description
TimeOpt: Evaluation of eccentricity-related amplitude modulation and bundling in paleoclimatedata, as in Meyers (2015).
Usage
timeOpt(dat,sedmin=0.5,sedmax=5,numsed=100,linLog=1,limit=T,fit=1,fitModPwr=T,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,detrend=T,output=0,title=NULL,genplot=T,check=T,verbose=T)
timeOpt 101
Arguments
dat Stratigraphic series for astrochronologic assessment. First column should bedepth or height (in meters), second column should be data value.
sedmin Minimum sedimentation rate for investigation (cm/ka).
sedmax Maximum sedimentation rate for investigation (cm/ka).
numsed Number of sedimentation rates to investigate in optimization grid.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log; default value is 1)
limit Limit evaluated sedimentation rates to region in which full target signal can berecovered? (T or F)
fit Test for (1) precession amplitude modulation or (2) short eccentricity amplitudemodulation?
fitModPwr Include the modulation periods in the spectral fit? (T or F)
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka)
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka)
roll Taner filter roll-off rate, in dB/octave.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with a first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
detrend Remove linear trend from data series? (T or F)
output Which results would you like to return to the console? (0) no output; (1) returnsedimentation rate grid, r^2_envelope, r^2_power, r^2_opt; (2) return optimaltime series, bandpassed series, envelope, reconstructed eccentricity model
title A character string (in quotes) specifying the title for the graphics window (op-tional)
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (T or F)
Details
TimeOpt is an astronomical testing algorithm for untuned (spatial) stratigraphic data. The algorithmidentifies the sedimentation rate(s) that simultaneously optimizes: (1) eccentricity amplitude mod-ulations within the precession band, and (2) the concentration of spectral power at specified targetastronomical periods.
For each temporal calibration investigated (i.e., sedimentation rate), the observed precession bandamplitude envelope is extracted using bandpass filtering and the Hilbert transform. The fit of theextracted precession envelope to the eccentricity periods is evaluated using a linear regression ontosine and cosine terms that reflect the five dominant eccentricity periods (~405.7, 130.7, 123.8, 98.9and 94.9 kyr); amplitude and phase of the eccentricity terms are not assigned, but are determined
102 timeOpt
during the linear model optimization. This approach is advantageous, as (1) the transfer func-tions associated with the climate and depositional systems can alter the amplitude and phase ofthe theoretical eccentricity terms (e.g, Laurin et al., 2005), and (2) the amplitude and phase of theeccentricity terms are unconstrained for deep-time investigations (>50 Ma). The quality of the "fit"is estimated by calculation of the correlation of the fitted eccentricity model time series to the ob-served precession band envelope (r^2_envelope), indicating the fraction of variance shared betweenthe model and envelope.
The concentration of power at the target astronomical periods is evaluated using a linear regressionof the temporally-calibrated series onto sine and cosine terms that reflect the dominant eccentricityand precession periods. As above, the amplitude and phase of each term is determined during thelinear model optimization, and the quality of the "fit" is estimated by calculation of the correlationof the fitted astronomical model series to the temporally-calibrated series (r^2_spectral).
The final measure of fit (r^2_opt) is determined as:
r^2_opt = r^2_envelope * r^2_spectral
which is simply the product of the fraction of variance shared between "model and envelope" and"model and time-calibrated data". This optimization approach identifies the sedimentation rate atwhich the precession envelope strongly expresses expected eccentricity modulation, while simul-taneously, spectral power is concentrated at the target astronomical periods. r^2_opt can take onvalues ranging from 0 to 1 (a perfect fit to the astronomical model), and provides a measure of over-all quality of the astronomically calibrated time series. A similar approach is applicable to evaluateshort eccentricity amplitude modulations. The statistical significance of the r^2_opt is determinedvia Monte Carlo simulation (see timeOptSim).
Value
if output = 1, a data frame containing the following will be returned: Sedimentation rate (cm/ka),r^2_envelope, r^2_spectral, r^2_opt
if output = 2, a data frame containing the following will be returned: Time (ka), tuned time series,bandpassed series, envelope, reconstructed model
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulation and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, 30, doi:10.1002/2015PA002850.
See Also
asm, eAsmTrack, testPrecession, timeOptPlot, and timeOptSim
Examples
# generate a test signal with precession and eccentricityex=etp(tmin=1,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE,genplot=FALSE,verbose=FALSE)# convert to meters with sedimentation rate of 2 cm/kyrex[1]<-ex[1]*0.02# evaluate precession modulationstimeOpt(ex,sedmin=0.5,sedmax=5,numsed=100,fit=1,output=0)
timeOptMCMC 103
# evaluate short eccentricity modulationstimeOpt(ex,sedmin=0.5,sedmax=5,numsed=100,fit=2,output=0)
timeOptMCMC TimeOptMCMC: Evaluation of eccentricity-related amplitude modu-lation and bundling in paleoclimate data ("TimeOpt"), with uncertain-ties via Markov-Chain Monte Carlo
Description
TimeOptMCMC: Evaluation of eccentricity-related amplitude modulation and bundling in paleo-climate data ("TimeOpt"; Meyers, 2015), with uncertainties on all fitting parameters via Markov-Chain Monte Carlo (MCMC). This function follows the approach of Meyers and Malinverno (2018).MCMC is implemented using the Metropolis-Hastings algorithm. Optimization is conducted uponthe sedimentation rate (constant within the study interval), the fundamental frequencies g1-g5, theprecession constant k, and four hyperparameters associated with the residuals from the spectral andenvelope fit. The priors for the k and g’s are Gaussian, while other parameters (sedrate, hyperpa-rameters) are uniform (uninformative).
Usage
timeOptMCMC(dat,iopt=1,sedmin=0.5,sedmax=5,sedstart=NULL,gAve=NULL,gSd=NULL,gstart=NULL,kAve=NULL,kSd=NULL,kstart=NULL,
rhomin=0,rhomax=0.9999,rhostart=NULL,sigmamin=NULL,sigmamax=NULL,sigmastart=NULL,ran=F,fit=1,ftol=0.01,roll=10^3,nsamples=1000,epsilon=NULL,test=F,burnin=-1,detrend=T,output=1,savefile=F,genplot=1,verbose=T)
Arguments
dat Stratigraphic series for astrochronologic assessment. First column should bedepth or height (in meters), second column should be data value.
iopt (1) fit power and envelope, (2) fit power only.
sedmin Minimum sedimentation rate for investigation (cm/ka).
sedmax Maximum sedimentation rate for investigation (cm/ka).
sedstart Initial sedimentation rate for MCMC search (cm/ka). Default is 0.5*(sedmin+sedmax).Alternatively, if set to negative number, a random value is selected from the priordistribution.
gAve Vector which contains the average values for the g1 through g5 fundamentalfrequencies (arcsec/year). Must be in the following order: g1,g2,g3,g4,g5.
gSd Vector which contains the standard deviation for the g1 through g5 fundamentalfrequencies (arcsec/year). Must be in the following order: g1,g2,g3,g4,g5.
gstart Vector which contains the initial values for the g1 through g5 fundamental fre-quencies (arcsec/year). Must be in the following order: g1,g2,g3,g4,g5. Defaultis 0.5*(gmin+gmax). Alternatively, if set to negative number, a random value isselected from the prior distribution.
104 timeOptMCMC
kAve Average value for the precession constant (arcsec/year).kSd Standard deviation for the precession constant (arcsec/year).kstart Initial value for the precession constant (arcsec/year). Default is 0.5*(kmin+kmax).
Alternatively, if set to negative number, a random value is selected from the priordistribution.
rhomin Minimum value for residual lag-1 autocorrelation (for both spectral and enve-lope fit). Default is 0.
rhomax Maximum value for residual lag-1 autocorrelation (for both spectral and enve-lope fit). Default is 0.9999
rhostart Initial value for residual lag-1 autocorrelation (for both spectral and envelopefit). Default 0.5. Alternatively, if set to negative number, a random value isselected from the prior distribution.
sigmamin Minimum value for residual sigma (for both spectral and envelope fit).sigmamax Maximum value for residual sigma (for both spectral and envelope fit).sigmastart Initial value for residual sigma (for both spectral and envelope fit). Default
0.5*(data standard deviation). Alternatively, if set to negative number, a randomvalue is selected from the prior distribution.
ran Would you like to randomly select the parameter for updating (T), or simultane-ously update all the parameters (F)?
fit Test for (1) precession amplitude modulation or (2) short eccentricity amplitudemodulation? Option 2 is not yet functional!
ftol Tolerance in cycles/ka used to define the precession bandpass. It is added tothe highest precession frequency, and subtracted from the lowest precession fre-quency, to define the half power points for the Taner bandpass filter.
roll Taner filter roll-off rate, in dB/octave.nsamples Number of candidate MCMC simluations to perform.epsilon Vector of dimension 11, which controls how large the jump is between each can-
didate value, e.g. sedimentation rate. For example, a value of 0.2 will yield max-imum jump +/- 10 percent of sedimentation rate range. The vector must be ar-ranged in the the following order: sedrate,k,g1,g2,g3,g4,g5,spec_rho,spec_sigma,env_rho,env_sigma.If NULL, all epsilon values will be assigned 0.2
test Activate epsilon testing mode? This option will assign all MCMC samples a log-likelihood of unity. This provides a diagnostic check to ensure that the appliedepsilon values are sampling the entire range of parameter values. (T or F)
burnin Threshold for detection of MCMC stability.detrend Remove linear trend from data series? (T or F)output Which results would you like to return to the console? (0) no output; (1) return
all MCMC candidatessavefile Save MCMC samples to file MCMCsamples.csv? (T or F). If true, results are
output after every 1000 iterations (last iterations will not be reported if you donot end on an even thousand!)
genplot Generate summary plots? (0= none; 1=display all summary plots; 2=also in-clude progress plot during iterations; 3=be quiet, but save all plots as pngs to theworking directory)
verbose Verbose output? (T or F)
timeOptMCMC 105
Details
TimeOpt is an astronomical testing algorithm for untuned (spatial) stratigraphic data. The algorithmidentifies the sedimentation rate(s) that simultaneously optimizes: (1) eccentricity amplitude mod-ulations within the precession band, and (2) the concentration of spectral power at specified targetastronomical periods.
This version of TimeOpt uses MCMC via Metropolis-Hastings to estimate the parameters and theiruncertainties. The priors for the k and g’s are Gaussian, while the other parameters (sedrate, hyper-parameters) are uniform (uninformative).
When ran=T, the following approach is used to select the parameter to modify:
0.25 probability of changing sedimentation rate
0.25 probability of changing k
0.30 probability of changing g1,g2,g3,g4,g5 (simultaneously)
0.10 probability of changing sigma_spec,rho_spec (simultaneously)
0.10 probability of changing sigma_env,rho_env (simultaneously)
This is motivated by sensitivity tests, and the fact that we are most interested in g, k and s; movingeach group of parameters (sedrate, k or g’s) has specific consequences we can isolate.
Here are some additional notes on the application of timeOptMCMC:
(1) Before conducting a timeOptMCMC analysis, run timeOpt to get a sense of the optimal sedi-mentation rate region(s).
(2) Make epsilon as large as you reasonably can, to maximize the chance of jumping betweenmodes. Think of epsilon as analogous to a diffusion coefficient. A good strategy is to run a coarseresolution analysis (large epsilon) to identify the optimum region, then use as small an epsilon aspossible to explore that optimum region. Note that larger epsilon yields less correlation in candi-dates. If you want to determine the time constant (thus number of independent samples) associatedwith a given epsilon, calculate the autocovariance function for accepted candidates (post-burnin).Decimation is useful for generating independent samples if desired.
(3) For greatest efficiency, the percentage of accepted candidates is typically expected to be be-tween 23-44 percent (see Gelman et al., 1996, "Efficient Metropolis jumping rules"). However, themultimodal nature of the parameter space may require smaller acceptance rates.
(4) To ensure that the MCMC algorithm is exploring the full parameter space, run an analysis with’test=T’. This option will accept all MCMC candidates. The histogram for each parameter valueshould approximate the prior distribution. If this is not the case, epsilon should be increased.
(5) It is expected that the MAP should be close to the mode when you have enough samples, al-though this is not guaranteed.
(6) There are different strategies for implementing the algorithm. One can run one long chain, orrun multiple short chains and combine.
(7) If you run a very long test chain, you can decimate to conduct a rarefaction analysis (of theparameters).
(8) For testing, it is recommended to run at least 3 very long chains. Ideally they should be longenough that you can’t tell the difference. Plot likelihood versus each candidate, and also sigma vseach candidate, for each run. This will allow identification of simulations that have gone into localminima.
106 timeOptPlot
(9) The following are useful estimates to consider: mean of candidate values (post-burnin), MAP,mode of kernel density estimate (post-burnin), 95 percent Credible Interval from kernel densityestimate (post-burnin).
(10) Keep in mind that a parabolic plot of log-likelihood vs. parameter value (quadratic in log-likelihood) indicates a Gaussian distribution.
For additional information see Meyers & Malinverno (2018), Meyers (2015), Tarantola (2005), andMalinverno & Briggs (2004).
References
A. Gelman et al., 1996, Efficient Metropolis jumping rules, Bayesian Statistics 5, p. 599-607.
A. Malinverno and V.A. Briggs, 2004, Expanded uncertainty quantification in inverse problems:Hierarchical Bayes and empirical bayes: Geophysics, 69, doi:10.1190/1.1778243.
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulation and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, 30, doi:10.1002/2015PA002850.
S.R. Meyers and A. Malinverno, 2018, Proterozoic Milankovitch cycles and the history of the solarsystem: Proceedings of the National Academy of Sciences, www.pnas.org/cgi/doi/10.1073/pnas.1717689115.
A. Tarantola, 2005, Inverse Problem Theory and Methods for Model Parameter Estimation, Societyfor Industrial and Applied Mathematics, 339 pages.
timeOptPlot TimeOptPlot: Generate summary figure for TimeOpt analyses
Description
TimeOptPlot: Generate summary figure for TimeOpt analyses.
Usage
timeOptPlot(dat=NULL,res1=NULL,res2=NULL,simres=NULL,fit=1,fitModPwr,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,xlab="Depth (m)",ylab="Proxy Value",fitR=NULL,verbose=T)
Arguments
dat Stratigraphic series used for astrochronologic assessment. First column shouldbe depth or height (in meters), second column should be data value.
res1 Data frame containing TimeOpt results: sedimentation rate grid, r^2_envelope,r^2_power, r^2_opt.
res2 Data frame containing the optimal-fitted time series, bandpassed series, enve-lope, and reconstructed eccentricity model.
simres Data frame containing the r^2_opt value for each Monte Carlo simulation.
fit Test for (1) precession amplitude modulation or (2) short eccentricity amplitudemodulation?
timeOptPlot 107
fitModPwr Include the modulation periods in the spectral fit? (T or F)
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka).
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka).
roll Taner filter roll-off rate, in dB/octave.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with a first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
xlab Label for the depth/height axis.
ylab Label for proxy variable evaluated.
fitR The r2_opt value at the optimal sedimentation rate.
verbose Verbose output? (T or F)
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulation and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, 30, doi:10.1002/2015PA002850.
See Also
asm, eAsmTrack, testPrecession, timeOpt, and timeOptSim
Examples
# generate a test signal with precession and eccentricityex=etp(tmin=1,tmax=1000,dt=1,pWt=1,oWt=0,eWt=1,esinw=TRUE,genplot=FALSE,verbose=FALSE)# convert to meters with sedimentation rate of 2 cm/kyrex[1]<-ex[1]*0.02# evaluate precession modulationsres1=timeOpt(ex,sedmin=0.5,sedmax=5,numsed=100,fit=1,output=1)res2=timeOpt(ex,sedmin=0.5,sedmax=5,numsed=100,fit=1,output=2)simres=timeOptSim(ex,sedrate=2,numsim=2000,fit=1,output=2)timeOptPlot(ex,res1,res2,simres,flow=0.035,fhigh=0.065,roll=10^3,targetE=c(405.6795,130.719,123.839,98.86307,94.87666),targetP=c(23.62069,22.31868,19.06768,18.91979),xlab="Depth (m)",ylab="Value",fitR=0.832,verbose=T)
108 timeOptSim
timeOptSim Monte Carlo simulation for TimeOpt
Description
Perform Monte Carlo AR1 simulations to evaluate significance of TimeOpt results, as in Meyers(2015).
Usage
timeOptSim(dat,numsim=2000,rho=NULL,sedrate=NULL,sedmin=0.5,sedmax=5,numsed=100,linLog=1,limit=T,fit=1,fitModPwr=T,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,detrend=T,ncores=2,output=0,genplot=T,check=T,verbose=T)
Arguments
dat Stratigraphic series for astrochronologic assessment. First column should bedepth or height (in meters), second column should be data value.
numsim Number of Monte Carlo AR1 simulations.
rho AR1 coefficient to use in simulations. By default this will be estimated from thestratigraphic series.
sedrate Sedimentation rate for investigation (cm/ka). This options is for compatibilitywith prior versions of timeOptSim. Please use sedmin, sedmax, numsed.
sedmin Minimum sedimentation rate for investigation (cm/ka).
sedmax Maximum sedimentation rate for investigation (cm/ka).
numsed Number of sedimentation rates to investigate in optimization grid.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log)
limit Limit evaluated sedimentation rates to region in which full target signal can berecovered? (T or F)
fit Test for (1) precession amplitude modulation or (2) short eccentricity amplitudemodulation?
fitModPwr Include the modulation periods in the spectral fit? (T or F)
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka)
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka)
roll Taner filter roll-off rate, in dB/octave.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with a first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
timeOptSim 109
detrend Remove linear trend from data series? (T or F)
ncores Number of cores to use for parallel processing. Must be >=2
output Which results would you like to return to console? (0) no output; (1) p-value;(2) simulation r2 results
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F). In general this shouldbe activated.
verbose Verbose output? (T or F)
Details
TimeOpt is an astronomical testing algorithm for untuned (spatial) stratigraphic data. The algo-rithm identifies the sedimentation rate(s) that simultaneously optimizes: (1) eccentricity amplitudemodulations within the precession band, and (2) the concentration of spectral power at specified tar-get astronomical periods. The statistical significance of the r^2_opt is determined via Monte Carlosimulation using timeOptSim.
The present version of timeOptSim improves upon the original significance testing method of Mey-ers (2015), by conducting simulations across the entire sedimentation grid. This approach morerigorously protects against inflation of the p-value due to multiple testing. Parallel processing hasbeen implemented to address the greater computational demand that is required.
See timeOpt for more information on the basic methodology.
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulations and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography.
See Also
asm, eAsm, eAsmTrack, testPrecession, timeOpt, and timeOptPlot
Examples
# generate a test signal with precession and eccentricityex=etp(tmin=1,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE,genplot=FALSE,verbose=FALSE)# convert to meters with sedimentation rate of 2 cm/kyrex[1]<-ex[1]*0.02# evaluate with timeOptSim. be patient, this may take a while to run.timeOptSim(ex,sedmin=0.5,sedmax=5,numsed=100)
110 timeOptTemplate
timeOptTemplate TimeOpt analysis using variable sedimentation models
Description
Evaluation of eccentricity-related amplitude modulations and bundling in paleoclimate data, as inMeyers (2015) and Meyers (2019), adapted to allow the evaluation of a wide range of variablesedimentation models, including: differential accumulation across bedding couplets, linear accu-mulation rate change, step changes in sedimentation rate, etc.
Usage
timeOptTemplate(dat,template=NULL,sedmin=0.5,sedmax=5,difmin=NULL,difmax=NULL,fac=NULL,numsed=50,linLog=1,limit=T,fit=1,fitModPwr=T,iopt=3,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,cormethod=1,detrend=T,detrendTemplate=F,flipTemplate=F,ncores=1,output=0,genplot=1,check=T,verbose=1)
Arguments
dat Stratigraphic series for astrochronologic assessment. First column should bedepth or height (in meters), second column should be data value.
template Instantaneous sedimentation rate template to fit. This represents a unitless pro-portional sedimentation rate history. Default model is a copy of dat, which willbe scaled for instantaneous accumulation optimization.
sedmin Minimum AVERAGE sedimentation rate for investigation (cm/ka).
sedmax Maximum AVERAGE sedimentation rate for investigation (cm/ka).
difmin Minimum instantaneous sedimentation rate to investigate (cm/ka).
difmax Maximum instantaneous sedimentation rate to investigate (cm/ka). By default,this is ignored, and fac is used.
fac Maximum instantaneous accumulation factor. Maximum rate is scaled to eachinvestigated sedrate as fac*sedrate. Default value of 5 is based on experimenta-tion. If larger than this, risk getting into local minimum during fit.
numsed Number of sedimentation rates to investigate in optimization grid.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log)
limit Limit evaluated sedimentation rates to region in which full target signal can berecovered? (T or F)
fit Test for (1) precession amplitude modulations or (2) short eccentricity amplitudemodulations?
fitModPwr Include the modulation periods in the spectral fit? (T or F)
iopt Optimize on (1) modulations, (2) spectral power, (3) modulations*spectral power
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka)
timeOptTemplate 111
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka)
roll Taner filter roll-off rate, in dB/octave. Default value is 10^3.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
cormethod Method used for calculation of correlation coefficient (1=Pearson, 2=Spearman)
detrend Remove linear trend from data series? (T or F)detrendTemplate
Remove linear trend from sedimentation rate template? (T or F)
flipTemplate Flip direction of sedimentation rate template? (T or F)
ncores Number of cores to use for parallel processing
output Which results you like to return to console? (0) no output; (1) return sedimenta-tion rate grid, r2; (2) return optimal time series, bandpassed series, Hilbert andfitted periods
genplot Generate summary plots? (0 = nothing, 1=summary plot, 2=progress + summaryplots)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (0 = nothing, 1=minimal, 2=everything)
Details
TimeOpt employs a probabilistic linear regression model framework to investigate amplitude mod-ulation and frequency ratios (bundling) in stratigraphic data, while simultaneously determining theoptimal time scale. This function further develops the method to optimize upon complex sedimen-tation templates. The approach is demonstrated below with a series of examples.
The statistical significance of the r^2_opt is determined via Monte Carlo simulation (see timeOpt-Sim). See timeOpt for more information on the basic methodology.
Value
if output = 1, a data frame containing the following will be returned: Sedimentation rate (cm/ka),r-squared value (r^2_envelope, r^2_spectra, or r^2_opt)
if output = 2, a data frame containing the following will be returned: Time (ka), tuned time series,bandpassed series, envelope, reconstructed model
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulations and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, v.30, 1625-1640.
S.R. Meyers, 2019, Cyclostratigraphy and the problem of astrochronologic testing: Earth-ScienceReviews v.190, 190-223.
112 timeOptTemplate
See Also
asm, eAsmTrack, testPrecession, timeOpt, timeOptSim, and timeOptTemplatePlot
Examples
## Not run:# EXAMPLE (1): Differential accumulation across bedding coupletsex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=diffAccum(ex,0.01,.05)ex2=linterp(ex2)# first with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)timeOptSim(ex2,sedmin=1,sedmax=4,numsed=100,numsim=2000)# then with the timeOptTemplate approachtimeOptTemplate(ex2,sedmin=1,sedmax=4,difmin=.5,difmax=6,numsed=100,ncores=2)timeOptTemplateSim(ex2,sedmin=1,sedmax=4,difmin=.5,difmax=6,numsed=100,numsim=1000,ncores=2)
# EXAMPLE (2): Linear sedimentation rate increaseex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=sedRamp(ex,srstart=0.01,srend=0.05)ex2=linterp(ex2)# first with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)# then with the timeOptTemplate approach# create linear model for input. the magnitude does not matter, it will be rescaled.# (it just needs to be a line)template=ex2; template[2]=ex2[1]timeOptTemplate(ex2,template=template,sedmin=1,sedmax=4,difmin=.5,difmax=6,numsed=100,ncores=2)# view optimization procedure (must set ncores=1)timeOptTemplate(ex2,template=template,sedmin=2.75,sedmax=3.25,difmin=.5,difmax=6,numsed=20,ncores=1,genplot=2)
# EXAMPLE (3): Step increase in sedimentation rate, from 1 cm/kyr to 2 cm/kyr at 7 meters depthex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=exex2[1]=ex[1]*.01ex2[141:201,1]=ex2[141:201,1]*2-7ex2=linterp(ex2)# first with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)# then with the timeOptTemplate approach# create step model for input. the magnitude does not matter, it will be rescaled.template=ex2; template[1:140,2]=1; template[141:261,2]=2timeOptTemplate(ex2,template=template,sedmin=1,sedmax=4,numsed=100,ncores=2)# view optimization procedure (must set ncores=1)timeOptTemplate(ex2,template=template,sedmin=1,sedmax=2,numsed=20,ncores=1,genplot=2)
# EXAMPLE (4): A record with a 100 kyr hiatus at 10 meters depthex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)
timeOptTemplatePlot 113
ex2=delPts(ex,del=101:121)# use a background sedimentation rate of 2 cm/kyrex2[1]=0:179*5*0.02# first evaluate the distorted record with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)# then with the timeOptTemplate approach# create a constant sedimentation rate template with possible hiatus of unknown# duration at 10 mtemplate=ex2; template[2]=10; template[101,2]=1timeOptTemplate(ex2,template=template,sedmin=1,sedmax=3,difmax=3,numsed=100,ncores=2)# now perform a finer grid search near the maximum, using power only# notice the oscillatory nature of the power fit.res=timeOptTemplate(ex2,template=template,sedmin=1.5,sedmax=2,difmax=3,numsed=100,ncores=2,iopt=2,output=2)
# compare true eccentricity to TimeOpt-derived eccentricitypl(2)plot(ex,type="l",main="True Eccentricity Series",xlab="True Time (kyr)",ylab="")plot(res[,1],res[,4],type="l",main="Black=TimeOpt precession AM; Red=TimeOpt eccentricity model",xlab="TimeOpt derived time (kyr)",ylab="")lines(res[,1],res[,5],col="red",lwd=2)
## End(Not run)
timeOptTemplatePlot TimeOptTemplatePlot: Generate summary figure for TimeOptTem-plate analyses
Description
TimeOptTemplatePlot: Generate summary figure for TimeOptTemplate analyses.
Usage
timeOptTemplatePlot(dat=NULL,template=NULL,detrend=T,detrendTemplate=F,flipTemplate=F,srMin=NULL,srMax=NULL,res1=NULL,simres=NULL,fit=1,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,xlab="Depth (m)",ylab="Proxy Value",fitR=NULL,output=0,verbose=T)
Arguments
dat Stratigraphic series used for astrochronologic assessment. First column shouldbe depth or height (in meters), second column should be data value.
template Instantaneous sedimentation rate template to fit. This represents a unitless pro-portional sedimentation rate history. Default model is a copy of dat, which willbe scaled for instantaneous accumulation optimization.
detrend Remove linear trend from data series? (T or F)detrendTemplate
Remove linear trend from sedimentation rate template? (T or F)
114 timeOptTemplatePlot
flipTemplate Flip direction of sedimentation rate template? (T or F)
srMin Minimum sedimentation rate for template
srMax Maximum sedimentation rate for template
res1 Data frame containing TimeOpt results: sedimentation rate grid, r^2_envelope,r^2_power, r^2_opt.
simres Data frame containing the r^2_opt value for each Monte Carlo simulation.
fit Test for (1) precession amplitude modulation or (2) short eccentricity amplitudemodulation?
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka).
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka).
roll Taner filter roll-off rate, in dB/octave.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with a first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
xlab Label for the depth/height axis.
ylab Label for proxy variable evaluated.
fitR The r2 value at the optimal sedimentation rate.
output Which results you like to return to console? (0) no output; (1) return sedimenta-tion rate grid, r2; (2) return optimal time series, bandpassed series, Hilbert andfitted periods
verbose Verbose output? (T or F)
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulations and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography, v.30, 1625-1640.
S.R. Meyers, 2019, Cyclostratigraphy and the problem of astrochronologic testing: Earth-ScienceReviews v. 190, 190-223.
See Also
asm, eAsmTrack, testPrecession, timeOpt, timeOptSim, and timeOptTemplate
timeOptTemplateSim 115
timeOptTemplateSim Simulations for timeOptTemplate
Description
Simulations for timeOptTemplate
Usage
timeOptTemplateSim(dat,template=NULL,corVal=NULL,numsim=2000,rho=NULL,sedmin=0.5,sedmax=5,difmin=NULL,difmax=NULL,fac=NULL,numsed=50,linLog=1,limit=T,fit=1,fitModPwr=T,iopt=3,flow=NULL,fhigh=NULL,roll=NULL,targetE=NULL,targetP=NULL,
cormethod=1,detrend=T,detrendTemplate=F,flipTemplate=F,ncores=1,output=0,genplot=T,check=T,verbose=T)
Arguments
dat Stratigraphic series for modulation assessment. First column should be depth orheight (in meters), second column should be data value.
template Instantaneous sedimentation rate template to fit. This represents a unitless pro-portional sedimentation rate history. Default template is a copy of dat, whichwill be scaled for instantaneous accumulation optimization.
corVal r2opt value for data. By default this will be calculated.
numsim Number of Monte Carlo AR1 simulations.
rho AR1 coefficient to use in simulations. By default this will be estimated from thestratigraphic series.
sedmin Minimum AVERAGE sedimentation rate for investigation (cm/ka).
sedmax Maximum AVERAGE sedimentation rate for investigation (cm/ka).
difmin Minimum instantaneous sedimentation rate to investigate (cm/ka).
difmax Maximum instantaneous sedimentation rate to investigate (cm/ka). By default,this is ignored, and fac is used.
fac Maximum instantaneous accumulation factor. Maximum rate is scaled to eachinvestigated sedrate as fac*sedrate. Default value of 5 is based on experimenta-tion. If larger than this, risk getting into local minimum during fit.
numsed Number of sedimentation rates to investigate in optimization grid.
linLog Use linear or logarithmic scaling for sedimentation rate grid spacing? (0=linear,1=log)
limit Limit evaluated sedimentation rates to region in which full target signal can berecovered? (T or F)
fit Test for (1) precession amplitude modulations or (2) short eccentricity amplitudemodulations? fit= 2 is not yet functional.
fitModPwr Include the modulation periods in the spectral fit? (T or F)
116 timeOptTemplateSim
iopt Optimize on (1) modulations, (2) power, (3) mod*power
flow Low frequency cut-off for Taner bandpass (half power point; in cycles/ka)
fhigh High frequency cut-off for Taner bandpass (half power point; in cycles/ka)
roll Taner filter roll-off rate, in dB/octave. Default value is 10^3.
targetE A vector of eccentricity periods to evaluate (in ka). These must be in order ofdecreasing period, with first value of 405 ka.
targetP A vector of precession periods to evaluate (in ka). These must be in order ofdecreasing period.
cormethod Method used for calculation of correlation coefficient (1=Pearson, 2=Spearman)
detrend Remove linear trend from data series? (T or F)
detrendTemplate
Remove linear trend from sedimentation rate template? (T or F)
flipTemplate Flip direction of sedimentation rate template? (T or F)
ncores Number of cores to use for parallel processing
output Which results you like to return to console? (0) no output; (1) return sedimenta-tion rate grid, p, r, r*p; (2) return optimal time series, bandpassed series, Hilbertand fitted periods
genplot Generate summary plots? (T or F)
check Conduct compliance checks before processing? (T or F) In general this shouldbe activated; the option is included for Monte Carlo simulation.
verbose Verbose output? (T or F)
Details
TimeOpt employs a probabilistic linear regression model framework to investigate amplitude mod-ulation and frequency ratios (bundling) in stratigraphic data, while simultaneously determining theoptimal time scale. This function further develops the method to optimize upon complex sedimen-tation templates. The approach is demonstrated below with a series of examples.
The statistical significance of the r^2_opt is determined via Monte Carlo simulation (see timeOpt-Sim). See timeOpt for more information on the basic methodology.
Value
QUESTION: is this correct?
if output = 1, a data frame containing the following will be returned: Sedimentation rate (cm/ka),r-squared value for instantaneous amplitude vs. fitted periods, r-squared value for fit to specifiedperiods, r-squared*r-squared.
if output = 2, a data frame containing the following will be returned: Time (ka), tuned time series,bandpassed series, instantaneous amplitude, fitted periods.
timeOptTemplateSim 117
References
S.R. Meyers, 2015, The evaluation of eccentricity-related amplitude modulations and bundling inpaleoclimate data: An inverse approach for astrochronologic testing and time scale optimization:Paleoceanography.
S.R. Meyers, 2019, Cyclostratigraphy and the problem of astrochronologic testing: Earth-ScienceReviews.
Examples
## Not run:
# EXAMPLE (1): Differential accumulation across bedding coupletsex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=diffAccum(ex,0.01,.05)ex2=linterp(ex2)# first with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)timeOptSim(ex2,sedmin=1,sedmax=4,numsed=100,numsim=2000)# then with the timeOptTemplate approachtimeOptTemplate(ex2,sedmin=1,sedmax=4,difmin=.5,difmax=6,numsed=100,ncores=2)timeOptTemplateSim(ex2,sedmin=1,sedmax=4,difmin=.5,difmax=6,numsed=100,numsim=1000,ncores=2)
# EXAMPLE (2): Linear sedimentation rate increaseex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=sedRamp(ex,srstart=0.01,srend=0.05)ex2=linterp(ex2)# first with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)# then with the timeOptTemplate approach# create linear model for input. the magnitude does not matter, it will be rescaled.# (it just needs to be a line)template=ex2; template[2]=ex2[1]timeOptTemplate(ex2,template=template,sedmin=1,sedmax=4,numsed=100,ncores=2)# view optimization proceduretimeOptTemplate(ex2,template=template,sedmin=2.75,sedmax=3.25,numsed=20,ncores=1,genplot=2)
# EXAMPLE (3): Step increase in sedimentation rate, from 1 cm/kyr to 2 cm/kyr at 7 meters depthex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=exex2[1]=ex[1]*.01ex2[141:201,1]=ex2[141:201,1]*2-7ex2=linterp(ex2)# first with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)# then with the timeOptTemplate approach# create step model for input. the magnitude does not matter, it will be rescaled.template=ex2; template[1:140,2]=1; template[141:261,2]=2timeOptTemplate(ex2,template=template,sedmin=1,sedmax=4,numsed=100,ncores=2)# view optimization procedure
118 tones
timeOptTemplate(ex2,template=template,sedmin=1,sedmax=2,numsed=20,ncores=1,genplot=2)
# EXAMPLE (4): A record with a 100 kyr hiatus at 10 meters depthex=etp(tmin=0,tmax=1000,dt=5,pWt=1,oWt=0,eWt=1,esinw=TRUE)ex2=delPts(ex,del=101:121)# use a background sedimentation rate of 2 cm/kyrex2[1]=0:179*5*0.02# first evaluate the distorted record with the nominal timeOpt approachtimeOpt(ex2,sedmin=1,sedmax=4,numsed=100)# then with the timeOptTemplate approach# create a constant sedimentation rate model with possible hiatus of unknown# duration at 10 mtemplate=ex2; template[2]=10; template[101,2]=1timeOptTemplate(ex2,template=template,sedmin=1,sedmax=3,difmax=3,numsed=100,ncores=2)# now perform a finer grid search near the maximum, using power only# notice the oscillatory nature of the power fit.res=timeOptTemplate(ex2,template=template,sedmin=1.5,sedmax=2,difmax=3,numsed=100,ncores=2,iopt=2,output=2)
# compare true eccentricity to TimeOpt-derived eccentricitypl(2)plot(ex,type="l",main="True Eccentricity Series",xlab="True Time (kyr)",ylab="")plot(res[,1],res[,4],type="l",main="Black=TimeOpt precession AM; Red=TimeOpt eccentricity model",xlab="TimeOpt derived time (kyr)",ylab="")lines(res[,1],res[,5],col="red",lwd=2)
## End(Not run)
tones Calculate all possible difference and combinations tones
Description
Determine all possible difference and combinations tones from a set of frequencies, and find theclosest one to a specified frequency
Usage
tones(a=NULL,freqs=NULL,f=T)
Arguments
a The frequency you are seeking to match, in cycles/ka.
freqs The vector of frequencies from which to calculate difference and combinationtones, in cycles/ka.
f Output results as frequencies (cycles/ka)? If false, will output results as periods(ka). (T or F)
traceFreq 119
traceFreq Frequency-domain minimal tuning: Use interactive graphical inter-face to trace frequency drift
Description
Frequency-domain minimal tuning: Use interactive graphical interface to trace frequency drift.
Usage
traceFreq(spec,color=2,h=6,w=4,ydir=1,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,ncolors=100,path=1,pl=0)
Arguments
spec Time-frequency spectral results to evaluate. Must have the following format:column 1=frequency; remaining columns (2 to n)=power, amplitude or proba-bility; titles for columns 2 to n must be the location (depth or height). Note thatthis format is ouput by function eha.
color Line color for tracing. 1 = transparent black; 2 = transparent white; 3 = trans-parent yellow
h Height of plot in inches.
w Width of plot in inches.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
xmin Minimum spatial frequency to plot.
xmax Maximum spatial frequency to plot.
ymin Minimum depth/height to plot.
ymax Maximum depth/height to plot.
ncolors Number of colors to use in plot.
path How do you want to represent the spatial frequency path?: 1=lines and points;2=lines; 3=points
pl An option for the color plots: 0=linear scale; 1=plot log of value, 2=normalizeto maximum value
See Also
eha and trackFreq
120 tracePeak
Examples
# Check to see if this is an interactive R session, for compliance with CRAN standards.# YOU CAN SKIP THE FOLLOWING LINE IF YOU ARE USING AN INTERACTIVE SESSION.if(interactive()) {
# Generate example series with 3 terms using function 'cycles'.# Then convert from time to space with sedimentation rate that increases from 1 to 5 cm/ka, using# function 'sedramp'.# Finally interpolate to median sampling interval using function 'linterp'.dat=linterp(sedRamp(cycles(freqs=c(1/100,1/40,1/20),start=1,end=2500,dt=5)))
# EHA anlaysis, output amplitude resultsout=eha(dat,output=3)
## Interactively track frequency driftfreq=traceFreq(out)
}
tracePeak A tool to interactively trace peak trajectories on plots
Description
A tool to interactively trace peak trajectories on plots, for results from such functions as eTimeOpt,EHA, eAsm.
Usage
tracePeak(dat,color=2,h=6,w=4,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,ydir=-1,ncolors=100,path=1)
Arguments
dat Data frame with results to evaluate. It must have the following format: column1=parameter to track (e.g., frequency, sedimentation rate, etc.; x-axis of plot);remaining columns (2 to n)=parameter to evaluate for peak identification (coloron plot); titles for columns 2 to n must be the location (depth/height/time; y-axisof plot). Note that this format is ouput by functions eha, eTimeOpt, eAsm.
color Line color for tracing. 1 = transparent black; 2 = transparent white; 3 = trans-parent yellow
h Height of plot in inches.
w Width of plot in inches.
ydir Direction for y-axis in plots (depth/height/time). -1 = values increase down-wards, 1 = values increase upwards.
trackFreq 121
xmin Minimum parameter value to plot.
xmax Maximum parameter value to plot.
ymin Minimum depth/height/time to plot.
ymax Maximum depth/height/time to plot.
ncolors Number of colors to use in plot.
path How do you want to represent the path?: 1=lines and points; 2=lines; 3=points
See Also
eha and eTimeOpt
trackFreq Frequency-domain minimal tuning: Use interactive graphical inter-face and sorting to track frequency drift
Description
Frequency-domain minimal tuning: Use interactive graphical interface and sorting algorithm totrack frequency drift.
Usage
trackFreq(spec,threshold=NULL,pick=T,fmin=NULL,fmax=NULL,dmin=NULL,dmax=NULL,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,h=6,w=4,ydir=1,ncolors=100,genplot=T,verbose=T)
Arguments
spec Time-frequency spectral results to evaluate. Must have the following format:column 1=frequency; remaining columns (2 to n)=power, amplitude or proba-bility; titles for columns 2 to n must be the location (depth or height). Note thatthis format is ouput by function eha.
threshold Threshold level for filtering peaks. By default all peak maxima reported.
pick Pick the peaks of interest using a graphical interface? (T or F). Only activated ifgenplot=T.
fmin Minimum frequency for analysis.
fmax Maximum frequency for analysis.
dmin Minimum depth/height for analysis. NOT ACTIVATED YET!
dmax Maximum depth/height for analysis. NOT ACTIVATED YET!
xmin Minimum frequency for PLOTTING.
xmax Maximum frequency for PLOTTING.
ymin Minimum depth/height for PLOTTING.
ymax Maximum depth/height for PLOTTING.
122 trackPeak
h Height of plot in inches.
w Width of plot in inches.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
ncolors Number of colors to use in plot.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
eha and traceFreq
Examples
# Check to see if this is an interactive R session, for compliance with CRAN standards.# YOU CAN SKIP THE FOLLOWING LINE IF YOU ARE USING AN INTERACTIVE SESSION.if(interactive()) {
# Generate example series with 3 terms using function 'cycles'.# Then convert from time to space with sedimentation rate that increases from 1 to 5 cm/ka, using# function 'sedramp'.# Finally interpolate to median sampling interval using function 'linterp'.dat=linterp(sedRamp(cycles(freqs=c(1/100,1/40,1/20),start=1,end=2500,dt=5)))
# EHA anlaysis, output probability resultsout=eha(dat,output=4)
## Isolate peaks with probability >= 0.8freq=trackFreq(out,0.8)
}
trackPeak A tool to interactively select points to track peak trajectories on plots
Description
A tool to interactively select points to track peak trajectories on plots, for results from functionssuch as eTimeOpt, EHA, eAsm.
Usage
trackPeak(dat,threshold=NULL,pick=T,minVal=NULL,maxVal=NULL,dmin=NULL,dmax=NULL,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,ydir=-1,h=6,w=4,ncolors=100,genplot=T,verbose=T)
trim 123
Arguments
dat Data frame with results to evaluate. It must have the following format: column1=parameter to track (e.g., frequency, sedimentation rate, etc.; x-axis of plot);remaining columns (2 to n)=parameter to evaluate for peak identification (coloron plot); titles for columns 2 to n must be the location (depth/height/time; y-axisof plot). Note that this format is ouput by functions eha, eTimeOpt, eAsm.
threshold Threshold level for filtering peaks. By default all peak maxima reported.
pick Pick the peaks of interest using a graphical interface? (T or F). Only activated ifgenplot=T.
minVal Minimum parameter value for analysis (e.g., frequency, sedimentation rate, etc.).
maxVal Maximum parameter value for analysis (e.g., frequency, sedimentation rate,etc.).
dmin Minimum depth/height/time for analysis. NOT ACTIVATED YET!
dmax Maximum depth/height/time for analysis. NOT ACTIVATED YET!
xmin Minimum parameter value for PLOTTING.
xmax Maximum parameter value for PLOTTING.
ymin Minimum depth/height/time for PLOTTING.
ymax Maximum depth/height/time for PLOTTING.
ydir Direction for y-axis in plots (depth or height). -1 = values increase downwards(slower plotting!), 1 = values increase upwards.
h Height of plot in inches.
w Width of plot in inches.
ncolors Number of colors to use in plot.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
eha and eTimeOpt
trim Remove outliers from stratigraphic series
Description
Automatically remove outliers from stratigraphic series, using ’boxplot’ algorithm.
Usage
trim(dat,c=1.5,genplot=T,verbose=T)
124 trimAT
Arguments
dat Stratigraphic series for outlier removal. First column should be location (e.g.,depth), second column should be data value.
c ’c’ defines the ’coef’ variable for boxplot.stats. For more information: ?box-plot.stats
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
delPts, idPts, iso and trimAT
trimAT Remove outliers from stratigraphic series
Description
Remove outliers from stratigraphic series, using specified threshold value.
Usage
trimAT(dat,thresh=0,dir=2,genplot=T,verbose=T)
Arguments
dat Stratigraphic series for outlier removal. First column should be location (e.g.,depth), second column should be data value.
thresh Threshold value for outlier detection.
dir Remove values (1) smaller than or (2) larger than this threshold?
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
See Also
delPts, idPts, iso and trim
trough 125
trough Identify minima of troughs in series, filter at desired threshold value
Description
Identify minima of troughs in any 1D or 2D series, filter at desired threshold value.
Usage
trough(dat,level,genplot=T,verbose=T)
Arguments
dat 1 or 2 dimensional series. If 2 dimesions, first column should be location (e.g.,depth), second column should be data value.
level Threshold level for filtering troughs. By default all trough minima reported.genplot Generate summary plots? (T or F)verbose Verbose output? (T or F)
Examples
ex=cycles(genplot=FALSE)trough(ex,level=-0.02)
tune Tune stratigraphic series
Description
Tune stratigraphic series from space to time, using specified control points
Usage
tune(dat,controlPts,extrapolate=F,genplot=T,check=T,verbose=T)
Arguments
dat Stratigraphic series for tuning. First column should be location (e.g., depth),second column should be data value.
controlPts Tuning control points. A data frame or matrix containing two columns: depth,time
extrapolate Extrapolate sedimentation rates above and below ’tuned’ interval? (T or F)genplot Generate summary plots? (T or F)check Conduct compliance checks before processing? (T or F) In general this should
be activated; the option is included for Monte Carlo simulation.verbose Verbose output? (T or F)
126 writeT
Examples
# generate example series with 3 precession terms using function 'cycles'ex1=cycles()
# then convert from time to space using a sedimentation rate that increases from 1 to 7 cm/kaex2=sedRamp(ex1,srstart=0.01,srend=0.07)
# assemble tuning control points (this is the depth-time map)controlPts=cbind(ex2[,1],ex1[,1])
# tune recordex3=tune(ex2,controlPts=controlPts)
writeCSV Write CSV file
Description
Write data frame as file with comma separated values
Usage
writeCSV(filename,output)
Arguments
filename Desired filename, in quotes: "result.csv"
output Data frame to write to file.
writeT Write tab-delimited file
Description
Write data frame as file with tab-delimited values
Usage
writeT(filename,output)
Arguments
filename Desired filename, in quotes: "result.tab"
output Data frame to write to file.
wtMean 127
wtMean Ar/Ar Geochronology: calculate weighted mean age, age uncertainty,and other associated statistics/plots (with interactive graphics for dataculling).
Description
The wtMean function will calculate weighted mean age, age uncertainty, and other helpful statis-tics/plots (with interactive graphics for data culling). The function includes the option to generateresults using the approach of IsoPlot 3.70 (Ludwig, 2008) or ArArCALC (Koppers, 2002).
Usage
wtMean(dat,sd=NULL,unc=1,lambda=5.463e-10,J=NULL,Jsd=NULL,CI=2,cull=-1,del=NULL,sort=1,output=F,idPts=T,size=NULL,unit=1,setAr=95,color="black",genplot=T,verbose=T)
Arguments
dat dat must contain one of the following: (1) a vector of dates for weighted meancalculation, (2) a matrix with two columns: date and uncertainty (one or twosigma), or (3) a matrix with six columns, as follows: date, date uncertainty(one or two sigma), K/Ca, %Ar40*, F, and F uncertainty (one or two sigma).NOTE: F is the ratio Ar40*/Ar39K (see Koppers, 2002). See "details" for moreinformation.
sd Vector of uncertainties associated with each date in ’dat’, as one or two sigma.This option is ignored if dat has more than one column
unc What is the uncertainty on your input dates? (1) one sigma, or (2) two sigma.DEFAULT is one sigma. This also applies to the F uncertainty, and the J-valueuncertainty (if specified)
lambda Total decay constant of K40, in units of 1/year. The default value is 5.463e-10/year (Min et al., 2000).
J Neutron fluence parameter
Jsd Uncertainty for J-value (neutron fluence parameter; as one or two sigma)
CI Which convention would you like to use for the 95% confidence intervals? (1)ISOPLOT (Ludwig, 2008), (2) ArArCALC (Koppers, 2002) (see below for de-tails)
cull Would you like select dates with a graphical interface? (0=no, 1=select pointsto retain, -1=select points to remove)
del A vector of indices indicating dates to remove from weighted mean calculation.If specified, this takes precedence over cull.
sort Sort dates? (0=no; 1=sort into increasing order; 2=sort into decreasing order)
output Return weighted mean results as new data frame? (T or F)
idPts Identify datum number on each point? (T or F)
128 wtMean
size Multiplicative factor to increase or decrease size of symbols and fonts.
unit The time unit for your results. (1) = Ma, (2) = Ka
setAr Set the %Ar40* level to be illustrated on the plot. The default is 95%.
color Color to use for symbols. Default is black.
genplot Generate summary plots? (T or F)
verbose Verbose output? (T or F)
Details
This function performs weighted mean age calculations, including estimation of age uncertainties,mean square weighted deviation, and probability of fit, following the approaches used in IsoPlot3.70 (Ludwig, 2008) and ArArCALC (Koppers, 2002).
The function accepts input in three formats:
(1) each date and its uncertainty can be entered as individual vectors (’dat’ and ’sd’).
(2) a two column matrix can be input as ’dat’, with each date (first column) and its uncertainty(second column).
(3) a six column matrix can be input as ’dat’, with each date, its uncertainty, the associated K/Cavalue, %Ar40*, F, and F uncertainty (one or two sigma). This option must be used if you wish tocalculate and include the uncertainty associated with J. The uncertainty is calculated and propagatedfollowing equation 18 of Koppers (2002).
The following plots are produced:
(1) A normal Q-Q plot for the dates (in essence this is the same as IsoPlot’s linearized probabilityplot).
(2) A cumulative Gaussian plot for the dates (a.k.a. cumulative probability plot). This is derived bysumming the individual normal distributions for each date.
(3) A plot of each date with its 2-sigma uncertainties.
In addition, K/Ca and Ar40* data are plotted if provided.
A NOTE regarding confidence intervals: There are two conventions that can be used to calculatethe confidence intervals, selected with the option ’CI’:
(1) ISOPLOT convention (Ludwig, 2008). When the probability of fit is >= 0.15, the confidenceinterval is based on 1.96*sigma. When the probability of fit is < 0.15, the confidence interval isbased on t*sigma*sqrt(MSWD).
(2) ArArCALC convention (Koppers, 2002). When MSWD <=1, the confidence interval is basedon 1.96*sigma. When MSWD > 1, the confidence interval is based on 1.96*sigma*sqrt(MSWD).
ADDITIONAL ADVICE: Use the function readMatrix to load your data in R (rather than the func-tion read).
References
A.A.P. Koppers, 2002, ArArCALC- software for 40Ar/39Ar age calculations: Computers & Geo-sciences, v. 28, p. 605-619.
K.R. Ludwig, 2008, User’s Manual for Isoplot 3.70: A Geochronological Toolkit for MicrosoftExcel: Berkeley Geochronology Center Special Publication No. 4, Berkeley, 77 p.
xplot 129
I. McDougall and T.M. Harrison, 1991, Geochronology and Thermochronology by the 40Ar/39ArMethod: Oxford University Press, New York, 269 pp.
K. Min, R. Mundil, P.R. Renne, and K. Ludwig, 2000, A test for systematic errors in 40Ar/39Argeochronology through comparison with U/Pb analysis of a 1.1-Ga rhyolite: Geochimica et Cos-mochimica Acta, v. 64, p. 73-98.
I. Wendt and C. Carl, 1991, The statistical distribution of the mean squared weighted deviation:Chemical Geology, v. 86, p. 275-285.
See Also
stepHeat
Examples
# Check to see if this is an interactive R session, for compliance with CRAN standards.# YOU CAN SKIP THE FOLLOWING LINE IF YOU ARE USING AN INTERACTIVE SESSION.if(interactive()) {
# Sample NE-08-01 Ar/Ar data from Meyers et al. (2012) supplementage <- c(93.66,94.75,94.6,94.22,86.87,94.64,94.34,94.03,93.56,93.85,88.55,93.45,93.84,
94.39,94.11,94.48,93.82,93.81,94.18,93.78,94.41,93.49,95.07,94.19)sd2<- c(5.83,4.10,8.78,2.5,8.86,3.37,4.63,3.18,8.35,5.73,4.23,2.56,2.3,1.7,3.1,2.78,
1.62,.92,.98,1.41,1.21,1.38,1.48,0.93)sd <- sd2/2wtMean(age,sd)
}
xplot Generate cross-plot with kernel density estimates on axes
Description
Generate a cross-plot with kernel density estimates on axes. If multiple data points are superposedin cross-plot, transparency of points reflects data density. Custom axes titles optional.
Usage
xplot(x,y,xlab=NULL,ylab=NULL,main=NULL,fill=T)
Arguments
x Variable 1
y Variable 2
xlab Label for the x-axis, in quotes
130 zoomIn
ylab Label for the y-axis, in quotes
main Label for the plot, in quotes
fill Use gray fill for density plots? (T or F)
Examples
# random numbers from a normal distributionex1<-rnorm(1000)# random numbers from an exponential distributionex2<-rexp(1000)
xplot(ex1,ex2)
zoomIn Dynamically explore cross-plot, zoom-in into specified region
Description
Dynamically explore cross-plot, zoom-in into specfied region. Accepts one dataframe/matrix withtwo columns, or two dataframes/vectors with one column.
Usage
zoomIn(dat1,dat2=NULL,ptsize=1,xmin=NULL,xmax=NULL,ymin=NULL,ymax=NULL,plotype=1,verbose=T)
Arguments
dat1 Data frame with one or two columns. If one column, dat2 must also be specified.
dat2 Data frame with one column.
ptsize Size of plotted points.
xmin Minimum x-value (column 1) to plot
xmax Maximum x-value (column 1) to plot
ymin Minimum y-value (column 2) to plot
ymax Maximum y-value (column 2) to plot
plotype Type of plot to generate: 1= points and lines, 2 = points, 3 = lines
verbose Verbose output? (T or F)
Index
∗Topic packageastrochron-package, 4
anchorTime, 6ar, 74ar1, 6ar1etp, 7arcsinT, 9, 22, 24, 48, 74, 75armaGen, 9asm, 10, 26, 27, 98, 100, 102, 107, 109, 112,
114astrochron (astrochron-package), 4astrochron-package, 4autoPlot, 12
bandpass, 13, 48, 66, 67, 74, 75, 93bergerPeriods, 14
calcPeriods, 15cb, 15clipIt, 16confAdjust, 17, 62, 96constantSedrate, 19cosTaper, 19, 24, 36, 38cycles, 20
delPts, 21, 40, 45, 124demean, 9, 22, 22, 24, 48, 74, 75detrend, 9, 22, 22, 24, 48, 74, 75diffAccum, 23divTrend, 9, 22, 23, 48, 74, 75dpssTaper, 20, 24, 36, 38
eAsm, 11, 25, 27, 109eAsmTrack, 11, 26, 26, 98, 100, 102, 107, 109,
112, 114eha, 26, 27, 27, 35, 44, 51, 54, 56, 58, 60, 119,
121–123eTimeOpt, 29, 32, 121, 123eTimeOptTrack, 30, 31etp, 8, 32
extract, 28, 34
flip, 35freq2sedrate, 35
gausTaper, 20, 24, 36, 38getColor, 36getData, 37getLaskar, 8, 33, 37
hannTaper, 20, 24, 36, 38headn, 39hilbert, 39
idPts, 21, 40, 45, 124imbrie, 41integratePower, 43iso, 21, 40, 45, 124
linage, 45linterp, 47logT, 9, 22, 24, 47, 74, 75lowpass, 13, 48, 66, 67, 74, 75, 93lowspec, 18, 28, 49, 54, 56, 58, 60, 62, 69, 96
makeNoise, 52modelA, 53mtm, 51, 53, 56, 58, 60, 62, 69, 96mtmAR, 28, 51, 54, 55, 58, 60mtmML96, 28, 51, 54, 56, 56, 60, 62, 96mtmPL, 59, 62, 96multiTest, 18, 60, 96mwCor, 62mwin, 64mwStats, 65
noKernel, 13, 48, 66, 67, 74, 75, 93noLow, 13, 48, 66, 67, 93
p.adjust, 62pad, 67
131
132 INDEX
peak, 68periodogram, 18, 28, 51, 54, 56, 58, 60, 62,
68, 96pl, 71plotEha, 72plS, 73prewhiteAR, 9, 13, 22, 24, 48, 66, 67, 73, 75,
93prewhiteAR1, 9, 13, 22, 24, 48, 66, 67, 74, 74,
93pwrLaw, 75pwrLawFit, 76
rankSeries, 77read, 77readMatrix, 78repl0, 79replEps, 80resample, 80rfbaseline, 51rmNA, 81runmed, 58
s, 81sedRamp, 82sedrate2time, 82slideCor, 83sortNave, 85spec.mtm, 18, 51, 54, 56, 58, 60stepHeat, 85, 129strats, 88surrogateCor, 84, 88surrogates, 89, 90synthStrat, 91
taner, 13, 48, 92testBackground, 18, 62, 94testPrecession, 11, 26, 96, 102, 107, 109,
112, 114testTilt, 98timeOpt, 11, 26, 30, 32, 98, 100, 100, 107,
109, 112, 114timeOptMCMC, 103timeOptPlot, 102, 106, 109timeOptSim, 11, 26, 30, 98, 100, 102, 107,
108, 112, 114timeOptTemplate, 110, 114timeOptTemplatePlot, 112, 113timeOptTemplateSim, 115
tones, 118traceFreq, 28, 119, 122tracePeak, 30, 120trackFreq, 28, 119, 121trackPeak, 30, 122trim, 21, 40, 45, 123, 124trimAT, 21, 40, 45, 124, 124trough, 125tune, 125
writeCSV, 126writeT, 126wtMean, 87, 127
xplot, 129
zoomIn, 130